Download - VMware View, EMC VNX, and Cisco UCS Proven Solution · PDF fileCISCO UNIFIED COMPUTING SYSTEMS . Proven Solution Guide . EMC Infrastructure for Virtual Desktops enabled by EMC VNX,

Proven Solution Guide

EMC GLOBAL SOLUTIONS

Abstract

This Proven Solution Guide provides a detailed summary of the tests performed to validate an EMC infrastructure for virtual desktops enabled by VMware View™ 4.5, with an EMC® VNX5700™ unified storage platform. This paper focuses on sizing and scalability, and highlights new features introduced in EMC VNX™, VMware vSphere™, and VMware® View. EMC unified storage uses advanced technologies like EMC FAST VP and EMC FAST Cache to optimize performance of the virtual desktop environment, helping to support service-level agreements.

May 2011

EMC INFRASTRUCTURE FOR VIRTUAL DESKTOPS ENABLED BY EMC VNX, VMWARE vSPHERE 4.1, VMWARE VIEW 4.5, VMWARE VIEW COMPOSER 2.5, AND CISCO UNIFIED COMPUTING SYSTEMS Proven Solution Guide

EMC Infrastructure for Virtual Desktops enabled by EMC VNX, VMware vSphere 4.1, VMware View 4.5, VMware View Composer 2.5, and

Cisco Unified Computing Systems—Proven Solution Guide

2

Copyright © 2011 EMC Corporation. All Rights Reserved.

EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice.

The information in this publication is provided “as is.” EMC Corporation makes no representations or warranties of any kind with respect to the information in this publication, and specifically disclaims implied warranties of merchantability or fitness for a particular purpose.

Use, copying, and distribution of any EMC software described in this publication requires an applicable software license.

For the most up-to-date listing of EMC product names, see EMC Corporation Trademarks on EMC.com.

VMware, VMware vCenter, VMware View, and VMware vSphere are registered trademarks or trademarks of VMware, Inc. in the United States and/or other jurisdictions. Iomega and IomegaWare are registered trademarks or trademarks of Iomega Corporation. All other trademarks used herein are the property of their respective owners.

Part Number H8197

Table of contents



3

Table of contents

1 Executive Summary .................................................................................... 12

Introduction to the VNX family of unified storage platforms ................................................ 12

Software suites available .......................................................................................................... 13

Software packs available ........................................................................................................... 13

Business case ..................................................................................................................... 14

Solution overview ............................................................................................................... 14

Key results and recommendations ...................................................................................... 15

2 Introduction ............................................................................................... 16

Introduction to the EMC VNX series ..................................................................................... 16

Document overview ............................................................................................................ 17

Purpose ..................................................................................................................................... 17

Scope ........................................................................................................................................ 17

Audience ................................................................................................................................... 17

Terminology .............................................................................................................................. 18

Technology overview ........................................................................................................... 19

Component list .......................................................................................................................... 19

EMC VNX platform ..................................................................................................................... 19

EMC Unisphere .......................................................................................................................... 19

EMC FAST VP ............................................................................................................................. 20

EMC FAST Cache ........................................................................................................................ 21 Block data compression ............................................................................................................ 21

Cisco Unified Computing System (UCS) ..................................................................................... 21

Solution diagram ................................................................................................................ 22

Configuration ...................................................................................................................... 23

Hardware resources ................................................................................................................... 23

Software resources .................................................................................................................... 24

3 Solution Infrastructure ................................................................................ 25

VMware View infrastructure ................................................................................................ 25

Introduction .............................................................................................................................. 25

VMware View components ......................................................................................................... 25

Hypervisor ................................................................................................................................. 26

VMware View Connection server ................................................................................................ 26

VMware vSphere vCenter/View Composer ................................................................................. 26

View Security server .................................................................................................................. 27

Table of contents



4

VMware View Transfer server ..................................................................................................... 27

Database server ........................................................................................................................ 27

VMware View Agent ................................................................................................................... 27

VMware View Client ................................................................................................................... 27

VMware View Admin Console .................................................................................................... 28

VMware View PowerCLI .............................................................................................................. 28

VMware ThinApp ....................................................................................................................... 28

VMware View virtual desktop infrastructure ........................................................................ 29

Introduction .............................................................................................................................. 29

Baseline .................................................................................................................................... 29

Processor .................................................................................................................................. 30

Memory ..................................................................................................................................... 30

Network ..................................................................................................................................... 31

Storage...................................................................................................................................... 31

vSphere 4.1 infrastructure .................................................................................................. 34

vCenter Server cluster ................................................................................................................ 34

Cisco Technology Overview ................................................................................................. 36

Overview ................................................................................................................................... 36

Cisco Unified Computing System (UCS) B-Series Blade Servers ................................................. 36

Cisco UCS 6100 Series Fabric Interconnects .............................................................................. 37

Cisco Nexus 7000 Series Switches ............................................................................................ 37

Cisco Nexus 1000V Series Switches, Cisco VN-Link technology ................................................. 37

Cisco MDS 9500 Series Multilayer Directors .............................................................................. 38

Windows infrastructure ....................................................................................................... 39

Introduction .............................................................................................................................. 39

Microsoft Active Directory .......................................................................................................... 39

Microsoft SQL Server ................................................................................................................. 39

DNS Server ................................................................................................................................ 39

DHCP Server .............................................................................................................................. 39

4 Network Design .......................................................................................... 40

Considerations ................................................................................................................... 40

Physical design considerations ................................................................................................. 40 Logical design considerations ................................................................................................... 40

Link aggregation ........................................................................................................................ 40

VNX for file network configuration ....................................................................................... 42

Data Mover ports ....................................................................................................................... 42

ESX network configuration .................................................................................................. 44

ESX NIC teaming ........................................................................................................................ 44

Port groups ................................................................................................................................ 44

Table of contents

5 EMC Infrastructure for Virtual Desktops enabled by EMC VNX, VMware vSphere 4.1, VMware View 4.5, VMware View Composer 2.5, and


Enterprise switch configuration .......................................................................................... 45

Cabling ...................................................................................................................................... 45

Server uplinks ........................................................................................................................... 45

Data Movers .............................................................................................................................. 45

Fibre Channel network configuration .................................................................................. 46

Introduction .............................................................................................................................. 46

Zone configuration .................................................................................................................... 47

5 Installation and Configuration ..................................................................... 48

Installation overview ........................................................................................................... 48

Installing VMware components ........................................................................................... 49

VMware View installation overview ............................................................................................ 49

VMware View setup ................................................................................................................... 49

VMware View desktop pool configuration .................................................................................. 49

PowerPath Virtual Edition .......................................................................................................... 56

Installing storage components ............................................................................................ 57

Create storage pools ................................................................................................................. 57

Enable FAST Cache .................................................................................................................... 59

Configure FAST VP ..................................................................................................................... 61

Configure VNX Home Directory ................................................................................................... 62

6 Testing and Validation ................................................................................ 64

Use case descriptions ......................................................................................................... 64

Use Case 1: FAST Cache with no dedicated replica LUN ............................................................. 64

Use Case 2: FAST Cache with dedicated replica LUNs ................................................................ 65

Use Case 3: A dedicated replica LUN with no FAST Cache .......................................................... 66

Boot storm scenario ............................................................................................................ 67

Overview ................................................................................................................................... 67

Use Case 1: FAST Cache with no dedicated replica LUN ............................................................. 67

Use Case 2: FAST Cache with dedicated replica LUNs ................................................................ 71

Use Case 3: A dedicated replica LUN with no FAST Cache .......................................................... 77

Antivirus scan scenario ....................................................................................................... 83

Overview ................................................................................................................................... 83

Use Case 1: FAST Cache with no dedicated replica LUN ............................................................. 83 Summary .......................................................................................................................................... 83 100-desktop antivirus scan ............................................................................................................... 84 200-desktop antivirus scan ............................................................................................................... 88 300-desktop antivirus scan ............................................................................................................... 93 500-desktop antivirus scan ............................................................................................................... 97

Use Case 2: With FAST Cache and a dedicated replica LUN ...................................................... 102

List of Tables



6

Summary ........................................................................................................................................ 102 100-desktop antivirus scan ............................................................................................................. 103 200-desktop antivirus scan ............................................................................................................. 108 300-desktop antivirus scan ............................................................................................................. 114 500-desktop antivirus scan ............................................................................................................. 120

Use Case 3: A dedicated replica LUN with no FAST Cache ........................................................ 126 Summary ........................................................................................................................................ 126 100-desktop antivirus scan ............................................................................................................. 127 200-descktop antivirus scan ........................................................................................................... 132 300-descktop antivirus scan ........................................................................................................... 138 500-desktop antivirus scan ............................................................................................................. 143

Antivirus scenario summary .................................................................................................... 148

Login VSI test scenario ...................................................................................................... 149

Overview ................................................................................................................................. 149

Use Case 1: with FAST Cache and no dedicated replica LUN .................................................... 149 With the Auto Tiering option enabled .............................................................................................. 149 With Performance Tiering option enabled ........................................................................................ 150

Use Case 2: with FAST Cache and with a dedicated replica LUNs ............................................. 156

Use Case 3: A dedicated replica LUN with no FAST Cache ........................................................ 162

7 Conclusion ............................................................................................... 165

Summary .......................................................................................................................... 165

Findings ............................................................................................................................ 165

References ........................................................................................................................ 166

White papers ........................................................................................................................... 166

Cisco documentation ............................................................................................................... 166

Other documentation .............................................................................................................. 166

List of Tables Table 1. VNX features .......................................................................................................................12 Table 2. Terminology ........................................................................................................................18 Table 3. Solution hardware ..............................................................................................................23 Table 4. Solution software ...............................................................................................................24 Table 5. Windows 7 configuration ...................................................................................................29 Table 6. Observed workload ............................................................................................................29 Table 7. Virtual machine per core ....................................................................................................30 Table 8. Required memory per host ................................................................................................31 Table 9. IOPS requirement and disks needed (multiple RAID scenarios) ........................................31 Table 10. Disks needed for RAID 5, RAID 10, RAID 6 ..........................................................................32 Table 11. Disks in storage tiering .......................................................................................................32 Table 12. Spindles used in this solution .............................................................................................33 Table 13. Port groups .........................................................................................................................44

List of Figures



List of Figures Figure 1. Unisphere Summary page ..................................................................................................20 Figure 2. Solution architecture ..........................................................................................................22 Figure 3. VMware components .........................................................................................................26 Figure 4. Linked clone ........................................................................................................................27 Figure 5. Linked clone virtual machine ..............................................................................................27 Figure 6. Cluster configuration from vCenter Server ........................................................................34 Figure 7. Virtual machines hosted on View-Cluster-2 .......................................................................35 Figure 8. Infrastructure cluster ..........................................................................................................35 Figure 9. Cisco Unified Computing System ........................................................................................36 Figure 10. SQL server databases ..........................................................................................................39 Figure 11. LACP configuration of the Data Mover ports .....................................................................41 Figure 12. VNX5700 Data Mover configuration ..................................................................................42 Figure 13. Virtual interface devices .....................................................................................................42 Figure 14. Interface properties ............................................................................................................43 Figure 15. vSwitch configuration in vCenter Server ............................................................................44 Figure 16. Data Mover port switch configuration ...............................................................................45 Figure 17. Zoning configuration ..........................................................................................................46 Figure 18. Zone configuration for the SAN B fabric .............................................................................47 Figure 19. Persistent automated desktop pools .................................................................................49 Figure 20. Select Automated Pool .......................................................................................................50 Figure 21. User assignment .................................................................................................................50 Figure 22. Select View Composer linked clones ..................................................................................51 Figure 23. Pool identification ..............................................................................................................52 Figure 24. Pool settings .......................................................................................................................52 Figure 25. Select Do not redirect Windows profile .............................................................................53 Figure 26. Provisioning settings ...........................................................................................................53 Figure 27. vCenter settings ..................................................................................................................54 Figure 28. Select the datastores for linked clone images....................................................................55 Figure 29. Guest customization ...........................................................................................................55 Figure 30. Verify your settings .............................................................................................................56 Figure 31. PowerPath as the owner for managing the path of block devices.....................................56 Figure 32. Thin LUNs created ..............................................................................................................57 Figure 33. Auto-Tiering ........................................................................................................................58 Figure 34. Enabling FAST Cache ...........................................................................................................59 Figure 35. FAST Cache configuration ...................................................................................................60 Figure 36. Configuring FAST VP ...........................................................................................................61 Figure 37. Configure the VNX Home Directory feature .......................................................................62 Figure 38. Sample Home Directory configuration ...............................................................................63 Figure 39. FAST Cache with no dedicated replica LUN ........................................................................64 Figure 40. FAST Cache with dedicated replica LUNs ...........................................................................65 Figure 41. Dedicated replica LUN with no FAST Cache .......................................................................66 Figure 42. LUN IOPS and response times for FAST Cache with no dedicated replica LUN .................67 Figure 43. Physical disk IOPS and response times ...............................................................................68 Figure 44. FAST Cache read and write operations ..............................................................................68 Figure 45. SP utilization during boot storm .........................................................................................69 Figure 46. Example boot-time SP utilization .......................................................................................69 Figure 47. ESX memory activity ...........................................................................................................70

List of Figures



8

Figure 48. ESX physical disk IOPS and guest latency ...........................................................................70 Figure 49. ESX VAAI statistics ..............................................................................................................71 Figure 50. LUN IOPS and response time ..............................................................................................71 Figure 51. Replica LUN IOPS and response times ................................................................................72 Figure 52. Physical disk IOPS and response time ................................................................................72 Figure 53. FAST Cache hit ratio ............................................................................................................73 Figure 54. SP utilization during the boot storm ..................................................................................73 Figure 55. Example ESX host physical CPU utilization .........................................................................74 Figure 56. ESX server memory .............................................................................................................74 Figure 57. ESX linked clone LUN IOPS and average guest latency ......................................................75 Figure 58. Linked clone ESX disk VAAI statistics ..................................................................................75 Figure 59. ESX replica disk IOPS and average guest latency................................................................76 Figure 60. ESX replica LUN VAAI statistics ...........................................................................................76 Figure 61. Linked clone LUN IOPS and response times .......................................................................77 Figure 62. Replica LUN IOPS and response times ................................................................................78 Figure 63. Physical disk IOPS and response times ...............................................................................78 Figure 64. Service processor (SP) utilization during boot storm .........................................................79 Figure 65. ESX server physical CPU utilization .....................................................................................79 Figure 66. ESX memory during boot storm .........................................................................................80 Figure 67. Linked clone ESX disk IOPS and average guest latency ......................................................80 Figure 68. Linked clone ESX disk VAAI .................................................................................................81 Figure 69. ESX replica disk IOPS and average guest latency................................................................81 Figure 70. ESX replica LUN VAAI statistics ...........................................................................................82 Figure 71. Antivirus scan summary with FAST Cash and no dedicated replica LUN ...........................83 Figure 72. Scan 100 LUN IOPS and response times .............................................................................84 Figure 73. Physical disk IOPS and response times ...............................................................................84 Figure 74. FAST Cache hit ratio ............................................................................................................85 Figure 75. 100-desktop antivirus scan SP utilization ...........................................................................85 Figure 76. ESX processor utilization ....................................................................................................86 Figure 77. ESX server memory utilization ............................................................................................86 Figure 78. 100-desktop antivirus scan ESX LUN IOPS and average guest latency ...............................87 Figure 79. Scan 100 - ESX VAAI statistics .............................................................................................87 Figure 80. Scan 100 - virtual machine disk IOPS and response times .................................................88 Figure 81. 200-desktop antivirus scan LUN IOPS and response times ................................................88 Figure 82. 200-desktop antivirus scan physical disk IOPS and response times ...................................89 Figure 83. 200-desktop antivirus scan FAST Cache Hit ratio ...............................................................89 Figure 84. 200 desktop antivirus scan SP utilization ...........................................................................90 Figure 85. ESX server processor utilization .........................................................................................90 Figure 86. 200-desktop antivirus scan ESX memory utilization ..........................................................91 Figure 87. 200-desktop antivirus scan ESX LUN IOPS and average guest latency ...............................91 Figure 88. ESX VAAI statistics ..............................................................................................................92 Figure 89. 200-desktop antivirus scan virtual machine disk IOPS and latency ...................................92 Figure 90. 300-desktop antivirus scan LUN IOPS and response times ................................................93 Figure 91. 300-desktop antivirus scan physical disk IOPS and response time ....................................93 Figure 92. 300-desktop antivirus scan FAST Cache hit ratio ...............................................................94 Figure 93. 300-desktop antivirus scan SP utilization ...........................................................................94 Figure 94. 300-desktop antivirus scan ESX CPU utilization .................................................................95 Figure 95. 300-desktop antivirus scan ESX memory utilization ..........................................................95 Figure 96. 300-desktop antivirus scan ESX LUN IOPS and average guest latency ...............................96

List of Figures



Figure 97. 300-desktop antivirus scan .................................................................................................96 Figure 98. 300-desktop antivirus scan virtual machine disk IOPS and latency ...................................97 Figure 99. 500-desktop antivirus scan LUN IOPS and response times ................................................97 Figure 100. 500-desktop antivirus scan physical disk IOPS and response times ...................................98 Figure 101. 500-desktop antivirus scan FAST Cache hit ratio ...............................................................98 Figure 102. 500-desktop antivirus scan service processor utilization ...................................................99 Figure 103. 500-desktop antivirus scan ESX server CPU utilization ......................................................99 Figure 104. 500-desktop antivirus scan ESX server memory utilization .............................................100 Figure 105. 500-desktop antivirus scan ESX disk IOPS and average guest latency .............................100 Figure 106. 500-desktop antivirus scan ESX VAAI statistics ................................................................101 Figure 107. 500-desktop antivirus scan virtual machine disk IOPS and latency .................................101 Figure 108. Antivirus scan summary with FAST Cache and a dedicated replica LUN .........................102 Figure 109. 100-desktop antivirus scan linked clone LUN IOPS and response times ..........................103 Figure 110. 100-desktop antivirus scan replica LUN IOPS and response times ..................................103 Figure 111. 100-desktop antivirus scan physical disk IOPS and response times .................................104 Figure 112. 100-desktop antivirus scan FAST Cache hit ratio .............................................................104 Figure 113. 100-desktop antivirus scan SP utilization .........................................................................105 Figure 114. 100-desktop antivirus scan ESX server CPU utilization ....................................................105 Figure 115. 100-desktop antivirus scan ESX memory utilization ........................................................106 Figure 116. 100-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency ........106 Figure 117. 100-desktop antivirus scan ESX VAAI statistics ................................................................107 Figure 118. 100-desktop antivirus scan ESX replica LUN IOPS and average guest latency .................107 Figure 119. 100-desktop antivirus scan ESX replica LUN VAAI statistics .............................................108 Figure 120. 200-desktop antivirus scan linked clone LUN IOPS and response times ..........................108 Figure 121. 200-desktop antivirus scan replica LUN IOPS and response times ..................................109 Figure 122. 200-desktop antivirus scan physical disk IOPS and response times .................................109 Figure 123. 200-desktop antivirus scan FAST Cache hit ratio .............................................................110 Figure 124. 200-desktop antivirus scan service processor utilization .................................................110 Figure 125. 200-desktop antivirus scan ESX CPU utilization ...............................................................111 Figure 126. 200-desktop antivirus scan ESX memory utilization ........................................................111 Figure 127. 200-desktop antivirus scan ESX LUN IOPS and average guest latency .............................112 Figure 128. 200-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................112 Figure 129. 200-desktop antivirus scan ESX replica LUN IOPS and average guest latency .................113 Figure 130. 200-desktop antivirus scan ESX replica LUN VAAI statistics .............................................113 Figure 131. 300-desktop antivirus scan linked clone LUN IOPS and response times ..........................114 Figure 132. 300-desktop antivirus scan replica LUN IOPS and response times ..................................114 Figure 133. 300-desktop antivirus scan physical disk IOPS and response times .................................115 Figure 134. 300-desktop antivirus scan FAST Cache hit ratio .............................................................115 Figure 135. 300-desktop antivirus scan service processor utilization .................................................116 Figure 136. 300-desktop antivirus scan ESX CPU utilization ...............................................................116 Figure 137. 300-desktop antivirus scan ESX memory utilization ........................................................117 Figure 138. 300-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency ........117 Figure 139. 300-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................118 Figure 140. 300-desktop antivirus scan ESX replica LUN IOPS and average guest latency .................119 Figure 141. 300-desktop antivirus scan VAAI statistics for the replica LUN .......................................119 Figure 142. 500-desktop antivirus scan LUN IOPS and response times ..............................................120 Figure 143. 500-desktop antivirus scan replica LUN IOPS and response times ..................................120 Figure 144. 500-desktop antivirus scan physical disk IOPS and response times .................................121 Figure 145. 500-desktop antivirus scan FAST Cache hit ratio .............................................................121

List of Figures



10

Figure 146. 500-desktop antivirus scan service processor utilization .................................................122 Figure 147. 500-desktop antivirus scan ESX server CPU utilization ....................................................122 Figure 148. 500-desktop antivirus scan ESX memory utilization ........................................................123 Figure 149. 500-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency ........123 Figure 150. 500-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................124 Figure 151. 500-desktop antivirus scan ESX replica LUN IOPS and average guest latency .................125 Figure 152. 500-desktop antivirus scan ESX replica LUN VAAI statistics .............................................125 Figure 153. Antivirus scan summary without FAST Cash but with a dedicated replica LUN ...............126 Figure 154. 100-desktop antivirus scan LUN IOPS and response times ..............................................127 Figure 155. 100-desktop antivirus scan replica LUN IOPS and response times ..................................127 Figure 156. 100-desktop antivirus scan physical disk IOPS and response times .................................128 Figure 157. 100-desktop antivirus scan service processor utilization .................................................128 Figure 158. 100-desktop antivirus scan ESX CPU utilization ...............................................................129 Figure 159. 100-desktop antivirus scan ESX memory utilization ........................................................129 Figure 160. 100-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency ........130 Figure 161. 100-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................130 Figure 162. 100-desktop antivirus scan ESX replica LUN IOPS and average guest latency .................131 Figure 163. 100-desktop antivirus scan ESX replica LUN VAAI statistics .............................................131 Figure 164. 200-desktop antivirus scan LUN IOPS and response time ................................................132 Figure 165. 200-desktop antivirus scan replica LUN IOPS and response times ..................................132 Figure 166. 200-desktop antivirus scan physical disk IOPS and response time ..................................133 Figure 167. 200-desktop antivirus scan service processor utilization .................................................133 Figure 168. 200-desktop antivirus scan ESX server CPU utilization ....................................................134 Figure 169. 200-desktop antivirus scan ESX memory utilization ........................................................134 Figure 170. 200-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency ........135 Figure 171. 200-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................136 Figure 172. 200-desktop antivirus scanESX replica LUN IOPS and average guest latency ..................137 Figure 173. 200-desktop antivirus scan ESX replica LUN VAAI statistics .............................................137 Figure 174. 300-desktop antivirus scan linked clone LUN IOPS and response times ..........................138 Figure 175. 300-desktop antivirus scan replica LUN IOPS and response times ..................................138 Figure 176. 300-desktop antivirus scan physical disk IOPS and response times .................................139 Figure 177. 300-desktop antivirus scan service processor utilization .................................................139 Figure 178. 300-desktop antivirus scan ESX CPU utilization ...............................................................140 Figure 179. 300-desktop antivirus scan ESX memory utilization ........................................................140 Figure 180. 300-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency ........141 Figure 181. 300-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................141 Figure 182. 300-desktop antivirus scan replica LUN IOPS and average guest latency ........................142 Figure 183. 300-desktop antivirus scan replica LUN VAAI statistics ...................................................142 Figure 184. 500-desktop antivirus scan linked clone LUN IOPS and response times ..........................143 Figure 185. 500-desktop antivirus scan replica LUN IOPS and resonse times ....................................143 Figure 186. 500-desktop antivirus scan physical disk IOPS and response times .................................144 Figure 187. 500-desktop antivirus scan service process utilization ....................................................144 Figure 188. 500-desktop antivirus scan ESX CPU utilization ...............................................................145 Figure 189. 500-desktop antivirus scan ESX memory utilization ........................................................145 Figure 190. 500-desktop antivirus scan ESX linked clone LUN IOPS and Average guest latency ........146 Figure 191. 500-desktop antivirus scan ESX linked clone LUN VAAI statistics ....................................146 Figure 192. 500-desktop antivirus scan ESX replica LUN IOPS and average guest latency .................147 Figure 193. 500-desktop antivirus scan ESX replica LUN VAAI statistics .............................................147 Figure 194. Average time to scan a single desktop .............................................................................148

List of Figures



Figure 195. Auto Tiering Login VSI results ...........................................................................................149 Figure 196. Performance Tiering Login VSI results ..............................................................................150 Figure 197. LUN IOPS and response times ..........................................................................................150 Figure 198. Physical disk IOPS and response time ..............................................................................151 Figure 199. FAST Cache read hit ratio .................................................................................................151 Figure 200. FAST Cache write hit ratio ................................................................................................152 Figure 201. FAST Cache hit ratio ..........................................................................................................152 Figure 202. Service processor utilization .............................................................................................153 Figure 203. ESX CPU utilization............................................................................................................153 Figure 204. ESX memory utilization .....................................................................................................154 Figure 205. ESX disk IOPS and average guest latency .........................................................................154 Figure 206. ESX disk VAAI statistics .....................................................................................................155 Figure 207. Virtual machine disk IOPS and latency .............................................................................155 Figure 208. Login VSI test results ........................................................................................................156 Figure 209. LUN IOPS and response times ..........................................................................................156 Figure 210. Replica LUN IOPS and response times ..............................................................................157 Figure 211. FAST Cache read hit ratio .................................................................................................157 Figure 212. FAST Cache write hit ratio ................................................................................................158 Figure 213. FAST Cache hit ratio ..........................................................................................................158 Figure 214. Service processor utilization .............................................................................................159 Figure 215. ESX server CPU utilization.................................................................................................159 Figure 216. ESX server memory utilization ..........................................................................................160 Figure 217. ESX linked clone LUN IOPS and average guest latency ....................................................160 Figure 218. ESX linked clone LUN VAAI statistics ................................................................................161 Figure 219. ESX replica LUN IOPS and average guest latency .............................................................161 Figure 220. ESX replica LUN VAAI statistics .........................................................................................162 Figure 221. Login VSI test results ........................................................................................................162 Figure 222. Linked clone LUN IOPS and response times .....................................................................163 Figure 223. Replica LUN IOPS and response times ..............................................................................163 Figure 224. Physical disk IOPS and response times .............................................................................164 Figure 225. Service process utilization ................................................................................................164

Chapter 1: Executive Summary



12

1 Executive Summary This chapter summarizes the proven solution described in this document and includes the following sections:

• Introduction to the VNX family of unified storage platforms

• Business case

• Solution overview

• Key results and recommendations

Introduction to the VNX family of unified storage platforms The EMC® VNX™ family delivers industry-leading innovation and enterprise capabilities for file, block, and object storage in a scalable, easy-to-use solution. This next-generation storage platform combines powerful and flexible hardware with advanced efficiency, management, and protection software to meet the demanding needs of today’s enterprises.

All of this is available in a choice of systems ranging from affordable entry-level solutions to high-performance, petabyte-capacity configurations servicing the most demanding application requirements. The VNXe™ series is purpose-built for the IT manager in entry-level environments, and the VNX series is designed to meet the high-performance, high-scalability requirements of midsize and large enterprises.

The VNX family includes two platform series:

• The VNX series, delivering leadership performance, efficiency, and simplicity for demanding virtual application environments that includes VNX7500™, VNX5700™, VNX5500™, VNX5300™, and VNX5100™

• The VNXe (entry) series with breakthrough simplicity for small and medium businesses that includes VNXe3300™ and VNXe3100™

Customers can benefit from new VNX features described in Table 1:

Table 1. VNX features

Features VNX series

VNXe series

Next-generation unified storage, optimized for virtualized applications

Capacity optimization features including compression, deduplication, thin provisioning, and application-centric copies

High availability, designed to deliver five 9s availability




Features VNX series

VNXe series

Automated tiering with Fully Automated Storage Tiering for Virtual Pools (FAST VP) and FAST Cache that can be optimized for the highest system performance and lowest storage cost simultaneously

Multiprotocol support for file and block protocols

Object access through Atmos™ Virtual Edition (Atmos VE)

Simplified management with EMC Unisphere™ for a single management framework for all NAS, SAN, and replication needs

Up to three times improvement in performance with the latest Intel multicore CPUs, optimized for Flash

Note: VNXe does not support block compression.

EMC provides a single, unified storage plug-in to view, provision, and manage storage resources from VMware vSphere™ across EMC Symmetrix®, VNX family, CLARiiON®, and Celerra® storage systems, helping users to simplify and speed up VMware storage management tasks.

The VNX family includes five new software suites and three new software packs, making it easier and simpler to attain the maximum overall benefits.

• FAST Suite—Automatically optimizes for the highest system performance and the lowest storage cost simultaneously (not available for the VNXe series or the VNX5100).

• Local Protection Suite—Practices safe data protection and repurposing (not applicable to the VNXe3100 as this functionality is provided at no additional cost as part of the base software).

• Remote Protection Suite—Protects data against localized failures, outages, and disasters.

• Application Protection Suite—Automates application copies and proves compliance.

• Security and Compliance Suite—Keeps data safe from changes, deletions, and malicious activity.

• Total Efficiency Pack—Includes all five software suites (not available for the VNX5100 and VNXe series).

• Total Protection Pack—Includes local, remote, and application protection suites (not applicable to the VNXe3100).

Software suites available

Software packs available




14

• Total Value Pack—Includes all three protection software suites and the Security and Compliance Suite (the VNX5100 and VNXe3100 exclusively support this package).

Business case Customers require a scalable, tiered, and highly available infrastructure on which to deploy their virtual desktop environment. There are several new technologies available to assist them in architecting a virtual desktop solution, but they need to know how to best use these technologies to maximize their investment, support service-level agreements, and reduce their desktop total cost of ownership (TCO).

The purpose of this solution is to build a replica of a common customer virtual desktop infrastructure (VDI) environment and to validate the environment for performance, scalability, and functionality. Customers will realize:

• Increased control and security of their global, mobile desktop environment, typically their most at-risk environment

• Better end-user productivity with a more consistent environment

• Simplified management with the environment contained in the data center

• Better support of service-level agreements and compliance initiatives

• Lower operational and maintenance costs

Solution overview This solution provides a detailed summary and characterization of the tests performed to validate an EMC infrastructure for virtual desktops enabled by VMware View 4.5 on an EMC VNX series platform. It involves building a 2,000-seat VMware View 4.5 environment on the EMC unified storage platform and integrates the new features of each of these systems to provide a compelling, cost-effective VDI platform.

This solution incorporates the following components and the EMC VNX5700 platform:

• 2,000 Microsoft Windows 7 virtual desktops

• VMware View Composer 2.5-based linked clones

• Storage tiering (SAS and NL-SAS)

• EMC FAST Cache

• EMC FAST VP

• Sizing and layout of the 2,000-seat VMware View 4.5 environment

• Multipathing and load balancing by EMC PowerPath®/VE

• User data on the CIFS share

• Redundant View Connection Manager




Key results and recommendations VMware View 4.5 virtualization technology meets user and IT needs, providing compelling advantages compared to traditional physical desktops and terminal services.

EMC VNX5700 brings flexibility to multiprotocol environments. With EMC unified storage, you can connect to multiple storage networks using NAS, iSCSI, and Fibre Channel SAN. EMC unified storage uses advanced technologies like EMC FAST VP and EMC FAST Cache to optimize performance for the virtual desktop environment. EMC unified storage supports vStorage APIs for Array Integration (VAAI), which were introduced in VMware vSphere 4.1. VAAI enables hosts to support more virtual machines per LUN and allows quicker virtual desktop provisioning. Zero paging recognition and transparent page sharing of the vSphere 4.1 feature helps you save memory and therefore allows you to host more virtual desktops per host.

Our team found the following key results during the testing of this solution:

• By using FAST Cache and VAAI, the time to concurrently boot all 2,000 desktops to a usable start was significantly reduced by 25 percent.

• By using a VAAI-enabled storage platform, we were able to store up to 512 virtual machines compared to 64 virtual machines per LUN.

• With VMware transparent page sharing, we observed memory savings of up to 92 GB on a host with 96 GB of RAM, and with less than 2 percent of it swapping to a FAST Cache-enabled LUN.

• Using Flash as FAST Cache for the read and write I/O operations reduced the number of spindles needed to support the required IOPS.

Chapter 2: Introduction



16

2 Introduction This chapter introduces the solution and its components, and includes the following sections:

• Introduction to the EMC VNX series

• Document overview

• Technology overview

• Solution diagram

• Configuration

Introduction to the EMC VNX series The EMC VNX series delivers uncompromising scalability and flexibility for the midtier storage users while providing market-leading simplicity and efficiency to minimize total cost of ownership. Customers can benefit from the new VNX features such as:

• Next-generation unified storage, optimized for virtualized applications

• Extended cache using Flash drives with FAST VP that can be optimized for the highest system performance and lowest storage cost simultaneously on both block and file.

• Multiprotocol support for file, block, and object with object access through Atmos Virtual Edition (Atmos VE).

• Simplified management with EMC Unisphere for a single management framework for all NAS, SAN, and replication needs.

• Up to three times improvement in performance with the latest Intel multicore CPUs, optimized for Flash.

• 6 Gb/s SAS back end with the latest drive technologies supported:

3.5-inch 100 GB and 200 GB Flash, 3.5-inch 300 GB, and 600 GB 15k or 10k rpm SAS, and 3.5-inch 2 TB 7.2k rpm NL-SAS

2.5-inch 300 GB and 600 GB 10k rpm SAS

• Expanded EMC UltraFlex™ I/O connectivity—Fibre Channel (FC), Internet Small Computer System Interface (iSCSI), Common Internet File System (CIFS), Network File System (NFS) including parallel NFS (pNFS), Multi-Path File System (MPFS), and Fibre Channel over Ethernet (FCoE) connectivity for converged networking over Ethernet.




Document overview This document provides a detailed summary of the tests performed to validate an EMC infrastructure for virtual desktops enabled by VMware View 4.5, with an EMC VNX5700 unified storage platform. It focuses on the sizing and scalability using features introduced in EMC’s VNX series, VMware vSphere 4.1, and VMware View 4.5. EMC unified storage uses advanced technologies like EMC FAST VP and EMC FAST Cache to optimize the performance of a virtual desktop environment, helping to support service-level agreements.

By integrating EMC VNX unified storage and the new features available in EMC’s VNX series and VMware View 4.5, desktop administrators are able to reduce costs by simplifying storage management and increase capacity utilization.

The purpose of this use case is to provide a virtualized solution for virtual desktops that is powered by VMware View 4.5, View Composer 2.5, VMware vSphere 4.1, EMC VNX series, EMC VNX FAST VP, VNX FAST Cache, and storage pools.

This solution includes all the attributes required to run this environment, such as hardware and software and the required VMware View configuration.

Information in this document can be used as the basis for a solution build, white paper, best practices document, or training. It can also be used by other EMC organizations (for example, the technical services or sales organizations) as the basis for producing documentation for a technical services or sales kit.

The paper contains the results of testing the EMC Infrastructure for Virtual Desktops Enabled by EMC VNX Series, VMware vSphere 4.1, VMware View 4.5, and VMware View Composer 2.5 solution.

Throughout this paper, we assume that you have some familiarity with the concepts and operations related to virtualization technologies and their use in information infrastructure.

This paper discusses multiple EMC products as well as those from other vendors. Some general configuration and operational procedures are outlined. However, for detailed product installation information, refer to the user documentation for those products.

The intended audience of this paper includes:

• Customers

• EMC partners

• Internal EMC personnel

Purpose

Scope

Audience




18

Table 2 provides terms frequently used in this paper.

Table 2. Terminology

Term Description

Block data compression

EMC unified storage introduces Block data compression, which allows customers to save and reclaim space anywhere in their production environment with no restrictions. This capability makes storage even more efficient by compressing data and reclaiming valuable storage capacity. Data compression works as a background task to minimize performance overhead. Block data compression also supports thin LUNs, and automatically migrates thick LUNs to thin during compression, freeing valuable storage capacity.

EMC FAST Cache This feature was introduced early with FLARE® release 30 and allows customers to use Flash drives as an expanded cache layer for the array. FAST Cache is an array-wide feature that you can enable for any LUN or storage pool. FAST Cache provides read and write access to the array.

EMC Fully Automated Storage Tiering for Virtual Pools (FAST VP)

EMC has enhanced its FAST technology to work at the sub-LUN level on both file and bock data. This feature works at the storage pool level, below the LUN abstraction. It supports scheduled migration of data to different storage tiers based on the performance requirements of individual 1 GB slices in a storage pool.

VMware Transparent Page Sharing

Transparent page sharing is a method by which redundant copies of memory pages are eliminated. Refer to http://kb.vmware.com/kb/1021095 for more information.

Linked clone A virtual desktop created by VMware View Composer from a writeable snapshot paired with a read-only replica of a master image.

Login VSI A third-party benchmarking tool developed by Login Consultants that simulates a real-world VDI workload by using an AutoIT script and determines the maximum system capacity based on the response time of the users.

Replica A read-only copy of a master image used to deploy linked clones.

Unisphere The centralized interface of the unified storage platforms. Unisphere includes integration with data protection services, provides built-in online access to key support tools, and is fully integrated with VMware.

VDI platform Virtual desktop infrastructure. The server computing model enabling desktop virtualization, encompassing the hardware and software system required to support the virtualized environment.

Terminology

http://kb.vmware.com/kb/1021095�




Term Description

Virtual desktop Desktop virtualization (sometimes called client virtualization), that separates a personal computer desktop environment from a physical machine using a client–server model of computing. The model stores the resulting "virtualized" desktop on a remote central server, instead of on the local storage of a remote client; therefore, when users work from their remote desktop client, all of the programs, applications, processes, and data used are kept and run centrally. This scenario allows users to access their desktops on any capable device, such as a traditional personal computer, notebook computer, smartphone, or thin client.

Technology overview

This section identifies and briefly describes the major components of the validated solution environment. The components are:

• EMC VNX platform

• EMC Unisphere

• EMC FAST Cache

• EMC FAST VP

• Block data compression

The EMC VNX platform brings flexibility to multiprotocol environments. With EMC unified storage, you can connect to multiple storage networks using NAS, iSCSI, and Fibre Channel SAN. EMC unified storage leverages advanced technologies like EMC FAST VP and EMC FAST Cache on VNX OE for block to optimize performance for the virtual desktop environment, helping support service-level agreements. EMC unified storage supports vStorage APIs for Array Integration (VAAI), which were introduced in VMware vSphere 4.1. VAAI enables quicker virtual desktop provisioning and start-up.

EMC Unisphere provides a flexible, integrated experience for managing CLARiiON, Celerra, and VNX platforms in a single pane of glass. This new approach to midtier storage management fosters simplicity, flexibility, and automation. Unisphere's unprecedented ease of use is reflected in intuitive task-based controls, customizable dashboards, and single-click access to real-time support tools and online customer communities.

Unisphere features include:

• Task-based navigation and controls that offer an intuitive, context-based approach to configuring storage, creating replicas, monitoring the environment, managing host connections, and accessing the Unisphere support ecosystem.

• A self-service Unisphere support ecosystem, accessible with one click from Unisphere, that provides users with quick access to real-time support tools, including live chat support, software downloads, product documentation, best

Component list

EMC VNX platform

EMC Unisphere




20

practices, FAQs, online communities, ordering spares, and submitting service requests.

• Customizable dashboard views and reporting capabilities that enable at-a-glance management by automatically presenting users with valuable information in terms of how they manage their storage. For example, customers can develop custom reports up to 18 times faster with EMC Unisphere.

• Common management provides a single sign-on and integrated experience for managing both block and file features.

Figure 1 provides an example of the Unisphere Summary page that gives administrators a wealth of detailed information on connected storage systems, from LUN pool and tiering summaries to physical capacity and RAID group information.

Figure 1. Unisphere Summary page

With EMC FAST VP, EMC has enhanced its FAST technology to be more automated with sub-LUN tiering and to support file as well as block. This feature works at the storage pool level, below the LUN abstraction. Where earlier versions of FAST VP operated above the LUN level, FAST VP now analyzes data patterns at a far more granular level. As an example, rather than move an 800 GB LUN to enterprise Flash drives, FAST VP now identifies and monitors the entire storage pool in 1 GB chunks. If data becomes active, then FAST VP automatically moves only these “hot” chunks to a higher tier like Flash. As data cools, FAST VP also correctly identifies which chunks to migrate to lower tiers and proactively moves them. With such granular tiering, it is now possible to reduce storage acquisition while at the same time improve performance and response time. In addition, because FAST VP is fully automated and policy-driven, no manual intervention is required to make this happen, so you save on operating costs as well.

EMC FAST VP




VNX FAST Cache, a part of the VNX FAST suite, enables Flash drives to be used as an expanded cache layer for the array. FAST Cache has array-wide features available for both file and block storage. FAST Cache works by examining 64 KB chunks of data in FAST Cache enabled objects on the array. Frequently accessed data is copied to the FAST Cache and subsequent accesses to that data chunk are serviced by FAST Cache. This allows immediate promotion of very active data to the Flash drives. This dramatically improves the response time for very active data and reduces the data hot spots that can occur within the LUN.

FAST Cache is an extended read/write cache that can absorb read-heavy activities such as boot storms and antivirus scans, and write-heavy workloads such as operating system patches and application updates.

EMC unified storage introduces block data compression, which allows customers to save and reclaim space anywhere in their production environment with no restrictions. This capability makes storage even more efficient by compressing data and reclaiming valuable storage capacity. Data Compression works as a background task to minimize performance overhead. Block data compression also supports thin LUNs, and automatically migrates thick LUNs to thin during compression, freeing valuable storage capacity.

Cisco UCS provides the computing platform purpose-built for virtualization, delivering a cohesive system that unites computing, networking, and storage access. Cisco UCS integrates a low-latency, lossless 10 Gigabit Ethernet unified network fabric with enterprise-class, x86-architecture servers that scale to the demands of virtualized desktop workloads, without sacrificing performance or application responsiveness. Cisco UCS Manager enables a stateless computing model employing Service Profile Templates that can scale up large pools of fully provisioned computing resources from “bare metal,” within a fraction of the time required by traditional server solutions.

EMC FAST Cache

Block data compression

Cisco Unified Computing System (UCS)




22

Solution diagram Figure 2 depicts the logical architecture of this solution.

Figure 2. Solution architecture




Configuration Table 3 lists the hardware used for this solution.

Table 3. Solution hardware

Hardware Quantity Configuration Notes

EMC VNX5700 1 DAEs configured with:

145 300 GB 15k rpm SAS disks

35 1 TB 7.2k near-line SAS disks

15 200 GB Flash drives

VNX shared storage providing block, file, FAST VP, and FAST Cache

Cisco UCS B200 server blade

16 Two quad-core Intel Xeon 5500 Family

48 GB RAM

Converged network adapter

Two UCS chassis, each hosting 8 blades

8 servers per vSphere cluster. Two clusters, each hosting 500 Windows 7 virtual machines.

Cisco UCS B200 server blade

8 Two quad-core Intel Xeon 5500 Family

96 GB RAM

B200-M2 blade with Xeon 5680 6 core proc

One UCS chassis of 8 blades. For one ESX® cluster hosting 1,000 Windows 7 virtual machines.

Intel server 2 Two quad-core Intel 5400 Family

32 GB RAM

Gigabit quad-port Intel VT

Infrastructure virtual machines (VMware vCenter™, DNS, DHCP, Active Directory, MS SQL Server, View Connection server and Replica Servers)

Cisco Nexus 7000 1 Infrastructure Ethernet switch

Cisco MDS 9509 2 For dual FC fabric

Cisco UCS Chassis

3 For 24 server blades, 6 UCS 2104XP IOMs

Windows 7 virtual desktops

Each 1 vCPU, 1.5 GB RAM, 20 GB VMDK, 1 (network interface card) NIC

Virtual desktops that are created for this solution

Hardware resources




24

Table 4 lists the software used with this solution.

Table 4. Solution software

Software Configuration Notes

EMC VNX5700

VNX OE for block

Release 31 Operating environment for the block

EMC VNX5700

VNX OE for file

Release 7.0 Operating environment for the file

VMware vSphere ESX 4.1 Build 260247 Server hypervisor

EMC PowerPath Virtual Edition 5.4 SP2 Multipathing and load balancing for block access.

VMware vCenter Server 4.1 vSphere Management Server

VMware View Manager 4.5 Software hosting virtual desktops

VMware View Composer 2.5 View component that uses linked clone technology to reduce storage size

Microsoft SQL Server 2005 Database that hosts the tables for VMware vCenter, View Composer, and View Events

Microsoft Windows 2008 R2 Operating system for the server environment

EMC Unisphere 1.0 Management tool for EMC VNX series

Microsoft Windows 7 64 bit RTM Operating system for the virtual desktops

VMware Tools 8.3.2 Enhancement tool for the virtual machine

Microsoft Office Office 2007 SP2 Used on the virtual desktops

Cisco UCS 1.2 Firmware, management software

Software resources

Chapter 3: Solution Infrastructure



3 Solution Infrastructure This chapter details the infrastructure of each component and includes the following sections:

• VMware View infrastructure

• VMware View virtual desktop infrastructure

• vSphere 4.1 infrastructure

• Windows infrastructure

VMware View infrastructure

VMware View delivers rich and personalized virtual desktops as a managed service from a virtualization platform built to deliver the entire desktop, including the operating system, applications, and user data. VMware View 4.5 provides centralized automated management of these components with increased control and cost savings. VMware View 4.5 improves business agility while providing a flexible high-performance desktop experience for end users across a variety of network conditions.

To provide a virtual desktop experience, VMware View uses various components, each with its own purpose. The components that make up the View environment are:

• Hypervisor

• VMware View Connection server

• VMware vSphere vCenter Server/View Composer

• VMware View Security server

• VMware View Transfer server

• Supported database server like Microsoft SQL Server

• VMware View Agent

• VMware View client

• VMware View Admin Console

• View PowerCLI

• ThinApp

Introduction

VMware View components




26

Figure 3 shows the VMware components described in the following sections.

Figure 3. VMware components

Hypervisor is used to host the virtual desktops. To get most of the features, we recommend that you use VMware vSphere 4. The vSphere 4 features such as vSphere API for Array Integration (VAAI), Memory Compression, and Ballooning help to host more virtual desktops on a host.

The VMware View Connection server hosts the LDAP directory and keeps the configuration information of the VMware View Desktop Pools, its associated virtual desktops, and VMware View. This data information can be replicated to other View Connection Replica servers. The Connection server also acts as a connection broker that maintains the desktop assignment. It supports a secure socket layer (SSL) connection to the desktop using remote desktop protocol (RDP) or protocol PC over IP (PCoIP). It also supports RSA® SecurID® two-factor authentication and smart card authentication.

The VMware vSphere vCenter server helps you manage your virtual machines and vSphere ESX hosts and provides high availability (HA) and Distributed Resource Scheduling (DRS) clusters. The VMware vCenter server hosts customization specification that permits cloned virtual machines to join the Active Directory (AD) domain. The View Composer service is installed on the vCenter server that provides storage savings by using linked clone technology to share the hard disk of parent virtual machines as shown in Figure 4.

Hypervisor

VMware View Connection server

VMware vSphere vCenter/View Composer




Figure 4. Linked clone

The operating system reads from the common read-only replica image and writes to the linked clone. Any unique data created by the virtual desktop is also stored in the linked clone. A logical representation of this relationship is shown in Figure 5.

Figure 5. Linked clone virtual machine

The View Security server is a different type of View Connection server. It supports two network interfaces—one to a private enterprise network and the other to the public network. It is typically used in a DMZ and enables users outside the organization to securely connect to their virtual desktops.

The VMware View Transfer server is another type of View Connection server that is required when you use the local mode feature. The Transfer server can use the CIFS share on VNX files to store the published image. The local mode allows users to work on a virtual desktop disconnected from the network and later synchronizes the changes with the View environment.

The VMware View supported database server is used to host the tables used by View Composer and can optionally store the VMware View events.

VMware View Agent is installed on the virtual desktop template and is deployed to all virtual desktops. It provides communication to the View Connection server and enables options for USB redirection, virtual printing, PCoIP server, and Smartcard over PCoIP.

VMware View Client software is used to connect to the virtual desktops using the connection broker. View Client allows users to print locally from their virtual desktop, and with the proper configuration, users can access USB devices locally.

View Security server

VMware View Transfer server

Database server

VMware View Agent

VMware View Client




28

VMware View Admin Console is a browser-based administration tool for VMware View and is hosted on the View Connection server.

VMware View PowerCLI provides the basic management of VMware View using Windows Powershell. It allows administrators to script some basic VMware View operations and can be used along with other Powershell scripts.

VMware ThinApp is an application virtualization product for enterprise desktop administrators and application owners. It enables rapid deployment of applications to physical and virtual desktops. ThinApp links the application, the ThinApp runtime, the virtual file system, and the virtual registry into a single package. The CIFS share on EMC VNX file can be used as a repository and to deploy the ThinApp to the virtual desktops.

VMware View Admin Console

VMware View PowerCLI

VMware ThinApp




VMware View virtual desktop infrastructure

This section provides information on how we designed our solution for hosting 2,000 users in a VMware View environment on EMC VNX series.

A Windows 7 desktop is loaded with the required applications and fine tuned for the virtual machine load. This includes removing unnecessary scheduled tasks, configurations, and services. For further details, refer to http://www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf

The configuration of the Windows 7 virtual machine is defined in Table 5.

Table 5. Windows 7 configuration

Device Configuration Notes

Processor 1 vCPU

Memory 1.5 GB

Hard disk 20 GB Replica on Flash, delta on SAS. No FAST Cache. No disposable disk. 64 K allocation unit.

Network interface card 1 vNIC

We ran a medium workload on a single virtual machine using Login VSI and observed the workload with a two-second interval during the execution of the test as described in Table 6.

Table 6. Observed workload

Login VSI Test - Medium Workload

OS User data disk Virtual machine

Read IOPS

Write IOPS

Total IOPS

Read IOPS

Write IOPS

Total IOPS

Active RAM

% Processor run time

Network total MB/s

Average 5.41 2.89 8.3 0.07 3.55 3.62 454 9 0.73

95th percentile 17.44 9.87 27.31 0 15.19 15.19 537.6 34.33 5.36

Max 402.56 67.77 470.33 14.36 43.08 57.44 537.6 99.45 18.01

Introduction

Baseline

http://www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf�

http://www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf�




30

The server used in this solution has two quad-core Intel Xeon 5500 processors. The average CPU load during the test is 9 percent. Therefore, we can run approximately 10 virtual machines per core. One host can run 2 × 4 × 10=80 virtual machines. The Intel Nehalem architecture is very efficient with hyper-threading and allows 50 to 80 percent more clients. This means it can run 1.5 × 80=120 to 1.8 × 80=144 virtual machines per host.

While using linked clones, up to eight hosts are allowed in a cluster. Leaving one node as failover capacity, with seven hosts, we can run 144×7=1008 virtual machines. One cluster can host 1,000 virtual desktops. Without considering the Intel Nehalem features, the cluster can host 80×7=560 virtual desktops. To host 2,000 virtual desktops, we need two to four clusters, which are about 128 to 256 processors in total. In a non-VDI environment, deploying 2,000 desktops would require 2,000 processors.

With hyper-threading, we are able to host 1,000 VMs per cluster and without hyper-threading, we are able to host only 500 VMs per cluster. Thus, we are able to double the number of hosts per cluster when using hyper-threading. In our solution, we use hyper-threading with three clusters; one with 1,000 users and other two with 500 users each. The 500-user cluster has extra room for processor intensive workloads.

Table 7 provides a summary of virtual machines per core.

Table 7. Virtual machine per core

Case Complete cluster Cluster with one node down

1000-user cluster 16 virtual machines per core 18 virtual machines per core

500-user cluster 8 virtual machines per core 9 virtual machines per core

One Windows 7 virtual machine is assigned 1.5 GB of memory. Without using VMware vSphere 4.1 features, it would require at least 9×8×1.5=108 GB to 18×8×1.5= 216 GB per host. VMware vSphere 4.1 provides features such as Transparent Page Sharing, ballooning, compression, recognition of zeroed pages, and memory compression that allows us to overcommit the memory to obtain a better consolidation ratio.

During the baseline workload, we observed about 540 MB used in active memory. The memory overhead was 179 MB; the hypervisor used 578 MB for the 48 GB host and 990 MB for the 96 GB host, and the service console memory was 561 MB. Based on this workload, we require 103 GB determined by the following calculation:

((9×8×(540+179) +578+561)/1024) = 52 GB to ((18×8×(540+179) +990+561)/1024) = 103 GB

VMware vSphere uses the above-mentioned features before it uses the swap memory. The FAST Cache on EMC’s VNX series storage platform does provide better response time compared to swapping the memory to SAS disks. Another option is to consider having a solid-state drive (SSD) on each host to host the vSwap. This may impact vMotion and also adds complexity to the environment. It is, therefore, advantageous to have the swap served by the FAST Cache on the EMC array.

Processor

Memory




Table 8 provides a summary of the memory required per host.

Table 8. Required memory per host

Case

RAM/host min required (Complete Cluster)

RAM / host min required (Cluster with one node down)

RAM/ host used on this solution

1,000 User Cluster 91 GB 103 GB 96 GB

500 User Cluster 46 GB 52 GB 48 GB

We used 1,536 GB in total for hosting 2,000 virtual desktops. In a typical case, 1.5 GB per desktop will not be available, instead 2 GB will be used, which would require 4,000 GB in total. Still, virtual desktops can provide better boot-up time compared to that of the traditional personal computer.

Based on the workload, we found one virtual machine requires approximately 18 Mb/s. So, a 100 Mb/s card can support five to six virtual machines per NIC, a 1 GB NIC can support 50 to 60 virtual machines per NIC, and a 10 GB NIC can support 500 to 600 virtual machines per NIC. The Converged Network Adapter (CNA) running at 50 percent bandwidth can support 250 to 300 virtual machines per CNA.

Note: This is just a rough estimate and we must always watch for the network load and look for the percent packet drops. If the value is high, check the network configuration and consider adding another NIC. In this solution, we used two CNAs per host to provide fault tolerance. For 2,000 virtual desktops, we used 2×8×3= 48 NICs. In a traditional desktop scenario, 2,000 desktops require 2,000 NICs.

The number of spindles required for hosting 2,000 user desktops is calculated using both the IOPS requirement and the capacity needed. Based on the workload, we observed 8.3 IOPS per virtual desktop on average. The maximum and 95th percentile is based on the time interval of the data. The sizing on the average IOPS can yield good performance for the virtual desktops operating in a steady state. However, this leaves insufficient headroom in the array to absorb high I/O peaks. To combat the issue of I/O storms, there should be two to three times the average to absorb that load. Table 9 details the IOPS requirement and Table 10 describes the disks needed by various RAID levels to meet that IOPS.

Table 9. IOPS requirement and disks needed (multiple RAID scenarios)

Item Value

Number of Windows 7 desktops 2,000

IOPS per Windows 7 virtual machines 9

Total host IOPS (HI) 18,000

% Read 65

% Write 35

Network

Storage




32

Table 10. Disks needed for RAID 5, RAID 10, RAID 6

Item Value

Total disk IOPS for RAID 5 (R5IO = HI × %R + HI×4×%W) 36,900

Number of SAS drives alone (R5IO/180) 205

Number of NL-SAS drives alone (R5IO/80) 462

Total disk IOPS for RAID 10 (R10IO = HI×%R + HI×2×%W) 24,300


Number of Flash drives alone (R10IO/2500) 10

Number of NL-SAS drives (R10IO/80) 304

Total Disk IOPS for RAID 6 (R6IO = HI×%R + HI×6×%W) 49,500


Number of NL-SAS drives alone (R6IO/80) 619

In keeping with the same IOPS and to increase performance or capacity, four SAS drives can be replaced with 9 NL-SAS, 125 SAS drives can be replaced with 9 Flash drives, and 125 NL-SAS drives can be replaced with 4 Flash drives. For a mix of 68 percent SAS, 1 percent NL-SAS, and 31 percent Flash, we need the disks as shown in Table 11 for various RAID options.

Table 11. Disks in storage tiering

Storage IOPS 36,900 (RAID 5) 49,500 (RAID 6) 24,300 (RAID 10)

SAS drive count 140 192 92

NL SAS drive count 5 12 4

Flash drive count 6 8 4

When considering the storage size of virtual desktops, VMware View Composer reduces the size required by using linked clone technology. Linked clones are dependent virtual machines linked to the replica virtual machine. A replica virtual machine is a thin provisioned copy of the master virtual machine. We deployed a 20 GB hard disk for the operating system to the master virtual machine. The files occupy 13 GB and, therefore, the replica virtual machine disk size is 13 GB.

In the desktop pool, we use a file share on the VNX array to host the user profile and data. A disposable disk that contains the temporary files and windows paging file is used to minimize the expansion of delta disks and, therefore, reduces the refresh frequency to the virtual machine.




The size of a virtual desktop is the size of the delta disk plus two times the memory size of the virtual machine plus 2 MB for internal disk plus the disposable disk and log size.

Considering 1 GB for the delta disk, it requires approximately 6 GB for one linked clone.

VMware View 4.5 supports 512 linked clones from a single replica. To host 500 virtual desktops, we need 3 TB. With the current VMFS version, the maximum size that is supported is 2 TB minus 512 bytes. This means that we need to split them into two datastores. To host 2,000 virtual desktops, we need eight datastores of 2 TB to allow additional space for growth. Therefore, we need 16 TB in total for the linked clones.

If linked clones are not used, it requires 25 GB per virtual machine in thick format or 18 GB per virtual machine using thin disks.

This solution uses 200 GB Flash, 300 GB SAS, and 2 TB NL-SAS disks. The usable raw capacities are 180 GB Flash, 268 GB SAS, and 1.8 TB NL-SAS.

With four Flash drives in RAID 10, 360 GB is dedicated for the replica with a RAID 5 mix of SAS and NL-SAS, and it gives 37 TB. With a RAID 10 mix of SAS and NL-SAS, we have 16 TB. With RAID 10, we use fewer spindles and that data does not grow much compared to user data.

With a dedicated datastore for the replica, the space that is required on the replica LUN is approximately 39 GB for three virtual desktop pools. Any data accessed three times in a given period normally resides in FAST Cache. To maximize the use of Flash, we elected to use it as FAST Cache. Table 12 describes the drives used in this solution.

Table 12. Spindles used in this solution

Drive Linked clone – RAID 10

User data- RAID 6

Hot spare FAST Cache

Flash 0 0 1 14

NL-SAS 4 32 2 NA

SAS 92 48 5 NA




34

vSphere 4.1 infrastructure

VMware vSphere 4.1 is the market-leading virtualization hypervisor used across thousands of IT environments around the world. VMware vSphere 4.1 can transform or virtualize computer hardware resources, including CPU, RAM, hard disk, and network controller, to create a fully functional virtual machine that runs its own operating system and applications just like a physical computer.

The high-availability features in VMware vSphere 4.1 along with VMware Distributed Resource Scheduler (DRS) and Storage vMotion® enable seamless migration of virtual desktops from one ESX server to another with minimal or no impact to customer usage.

Figure 6 shows the cluster configuration from vCenter Server. The clusters View-Cluster-1 and View-Cluster-2 host 500 virtual desktops, while View-Cluster-5 hosts 1,000 virtual desktops.

Figure 6. Cluster configuration from vCenter Server

vCenter Server cluster




Figure 7 shows the virtual machines hosted on View-Cluster-2, its failover capacity, and its memory utilization.

Figure 7. Virtual machines hosted on View-Cluster-2

The infrastructure cluster hosts the SQL Server, vCenter, domain controller, and View Connection servers as shown in Figure 8.

Figure 8. Infrastructure cluster




36

Cisco Technology Overview

Figure 9 displays the Cisco UCS components described in this section.

Figure 9. Cisco Unified Computing System

Cisco UCS B-Series Blade Servers are designed for compatibility, performance, energy efficiency, large memory footprints, manageability, and unified I/O connectivity:

• Compatibility: Cisco UCS B-Series Blade Servers are designed around multicore Intel Xeon 5500, 5600, 6500 and 7500 Series processors, DDR3 memory, and an I/O bridge. Each blade server's front panel provides direct access for video, two USB, and console connections.

• Performance: Cisco's blade servers use the Intel Xeon next-generation server processors, which deliver intelligent performance, automated energy efficiency, and flexible virtualization. Intel Turbo Boost Technology automatically boosts processing power through increased frequency and use of hyper-threading to deliver high performance when workloads demand and thermal conditions permit. Intel Virtualization Technology provides best-in-class support for virtualized environments, including hardware support for direct connections between virtual machines and physical I/O devices.

• Energy efficiency: Most workloads vary over time. Some workloads are bursty on a moment-by-moment basis, while others have predictable daily, weekly, or monthly cycles. Intel Intelligent Power Technology monitors the CPU utilization and automatically reduces energy consumption by putting processor cores into a low-power state based on real-time workload characteristics.

Overview

Cisco Unified Computing System (UCS) B-Series Blade Servers




• Large-memory-footprint support: As each processor generation delivers even more power to applications, the demand for memory capacity to balance CPU performance increases as well. The widespread use of virtualization increases memory demands even further due to the need to run multiple OS instances on the same server. Cisco blade servers with Cisco Extended Memory Technology can support up to 384 GB per blade.

• Manageability: The Cisco Unified Computing System is managed as a cohesive system. Blade servers are designed to be configured and managed by Cisco UCS Manager, which can access and update blade firmware, BIOS settings, and RAID controller settings from the parent Cisco UCS 6100 Series Fabric Interconnect. Environmental parameters are also monitored by Cisco UCS Manager, reducing the number of points of management.

• Unified I/O: Cisco UCS B-Series Blade Servers are designed to support up to two network adapters. This design can reduce the number of adapters, cables, and access-layer switches by as much as half because it eliminates the need for multiple parallel infrastructure for both LAN and SAN at the server, chassis, and rack levels. This design results in reduced capital and operating expenses through lower administrative overhead and power and cooling requirements.

A core part of the Cisco Unified Computing System, the Cisco UCS 6100 Series Fabric Interconnects provide both network connectivity and management capabilities to all attached blades and chassis. The Cisco UCS 6100 Series offers line-rate, low-latency, lossless 10 Gigabit Ethernet and FCoE functions. The interconnects provide the management and communication backbone for the Cisco UCS B-Series Blades and UCS 5100 Series Blade Server Chassis.

The Cisco Nexus 7000 Series offers an end-to-end solution for data center core, aggregation, and high-density end-of-row and top-of-rack server connectivity in a single platform. The Cisco Nexus 7000 Series platform is run by Cisco NX-OS software. It is specifically designed for the most mission-critical place in the network, the data center.

Cisco Nexus 1000V Series Switches are virtual machine access switches that are an intelligent software switch implementation based on IEEE 802.1Q standard for VMware vSphere environments running the Cisco NX-OS Software operating system. Operating inside the VMware ESX hypervisor, the Cisco Nexus 1000V Series supports Cisco VN-Link server virtualization technology to provide:

• Policy-based virtual machine connectivity

• Mobile virtual machine security and network policy

• An undisruptive operational model for server virtualization and networking teams

Cisco UCS 6100 Series Fabric Interconnects

Cisco Nexus 7000 Series Switches

Cisco Nexus 1000V Series Switches, Cisco VN-Link technology




38

The Cisco MDS 9500 Series Multilayer Director layers a broad set of intelligent features onto a high-performance, open-protocol switch fabric. Addressing the stringent requirements of large data center storage environments, it provides high availability, security, scalability, ease of management, and transparent integration of new technologies.

Cisco MDS 9500 Series Multilayer Directors




Windows infrastructure Microsoft Windows provides the infrastructure used to support the virtual desktops and includes the following components:

• Microsoft Active Directory

• Microsoft SQL Server

• DNS Server

• DHCP Server

The Windows domain controller runs the Active Directory service that provides the framework to manage and support the virtual desktop environment. Active Directory provides several functions to help you:

• Manage the identities of users and their information

• Apply group policy objects

• Deploy software and updates

Microsoft SQL Server is a relational database management system (RDBMS). SQL Server 2008 is used to provide the required databases to vCenter Server, View Composer, and View Events as shown in Figure 10.

Figure 10. SQL server databases

DNS is the backbone of Active Directory and provides the primary name resolution mechanism of Windows servers and clients. In this solution, the DNS role is enabled on the domain controller.

The DHCP Server provides the IP address, DNS Server name, gateway address, and other information to the virtual desktops. In this solution, we enabled the DHCP role on the domain controller.

Introduction

Microsoft Active Directory

Microsoft SQL Server

DNS Server

DHCP Server

Chapter 4: Network Design



40

4 Network Design This chapter describes the network design used in this solution and contains the following sections:

• Considerations

• VNX for file network configuration

• Enterprise switch configuration

• Fibre Channel network configuration

Considerations

EMC recommends that switches support gigabit Ethernet (GbE) connections and Link Aggregation Control Protocol (LACP), and the ports on switches support copper-based media.

This validated solution uses virtual local area networks (VLANs) to segregate network traffic of various types to improve throughput, manageability, application separation, high availability, and security.

The IP scheme for the virtual desktop network must be designed so that there are enough IP addresses in one or more subnets for the DHCP Server to assign them to each virtual desktop.

VNX platforms provide network high availability or redundancy by using link aggregation. This is one of the methods used to address the problem of link or switch failure.

Link aggregation is a high-availability feature that enables multiple active Ethernet connections to appear as a single link with a single MAC address and potentially multiple IP addresses.

In this solution, LACP is configured on VNX, which combines eight GbE ports into a single virtual device. If a link is lost in the Ethernet port, the link fails over to another port. All network traffic is distributed across the active links. Figure 11 shows the LACP configuration of the Data Mover ports on the Ethernet switch.

Physical design considerations

Logical design considerations

Link aggregation




Figure 11. LACP configuration of the Data Mover ports




42

VNX for file network configuration

VNX5700 consists of two Data Movers, which can be configured in an active/active or active/passive configuration. In this solution, the Data Movers operate in the active/passive mode. In the active/passive configuration, the passive Data Mover serves as a failover device for the active Data Mover.

The VNX5700 Data Mover is configured with two UltraFlex™ I/O modules with each consisting of four 1 Gb interfaces. It is configured to use LACP with all Data Mover ports as shown in Figure 12.

Figure 12. VNX5700 Data Mover configuration

The lacp0 device was used to support virtual machine traffic, home folder access, and external access for roaming profiles. The virtual interface devices were created on the same LACP for each VLAN that requires access to the Data Mover interfaces as shown in Figure 13.

Figure 13. Virtual interface devices

Data Mover ports




Figure 14 shows the properties of a single interface where the VLAN ID and Maximum Transfer Unit (MTU) are set.

Figure 14. Interface properties




44

ESX network configuration

All network interfaces in this solution use 10 GbE connections. The server Ethernet ports on the switch are configured as trunk ports and use VLAN tagging at the port group to separate the network traffic between various port groups. Figure 15 shows the vSwitch configuration in vCenter Server.

Figure 15. vSwitch configuration in vCenter Server

Table 13 lists the configured port groups.

Table 13. Port groups

Configured port groups Used to

Virtual machine network Provide external access for administrative virtual machines

Service Console Manage public network administration traffic

Desktop-Network Provide a network connection for virtual desktops and LAN traffic

ESX NIC teaming

Port groups




Enterprise switch configuration In this solution, we spread the ESX server and VNX Data Mover cabling evenly across two line cards to provide redundancy and load balancing of the network traffic.

The server uplinks to the switch are configured in a port channel group to increase the utilization of server network resources and to provide redundancy. The vSwitches are configured to load balance the network traffic on the originating port ID.

We used the following configuration for one of the server ports in this solution:

switchport switchport trunk encapsulation dot1q switchport mode trunk no ip address

spanning-tree portfast trunk

The network ports for each VNX5700 Data Mover are connected to the Ethernet switch. The ports are configured with LACP, which provides redundancy in case of a NIC or port failure.

Figure 16 shows an example of the switch configuration for one of the Data Mover ports.

Figure 16. Data Mover port switch configuration

Cabling

Server uplinks

Data Movers




46

Fibre Channel network configuration

Enterprise-class FC switches are used to provide the storage network for this solution. The switches are configured in a SAN A/SAN B configuration to provide fully redundant fabrics.

Each server has a single connection to each fabric to provide load-balancing and failover capabilities. Each storage processor has two links to the SAN fabrics for a total of four available front-end ports. The zoning is configured so that each server has four available paths to the storage array. Figure 17 confirms that information from the vCenter interface.

Figure 17. Zoning configuration

Introduction




This solution uses single initiator and multiple target zoning. Each server initiator is zoned to two storage targets on the array. Figure 18 shows the zone configuration for the SAN B fabric.

Figure 18. Zone configuration for the SAN B fabric

Zone configuration

Chapter 5: Installation and Configuration



48

5 Installation and Configuration This chapter describes how to install and configure this solution and includes the following sections:

• Installation overview

• Installing VMware components

• Installing Storage components

Installation overview

This section describes how to configure both the VMware and storage components in this solution, including:

• Desktop pools

• Storage pools

• FAST Cache

• Auto-tiering (FAST VP)

• VNX Home Directory

• PowerPath/VE

The installation and configuration steps for the following components are available on the VMware website www.vmware.com:

• VMware View Connection Server

• VMware View Composer 2.5

• VMware ESX 4.1

• VMware vSphere 4.1

The installation and configuration of the following components are not covered:

• Microsoft System Center Configuration Manager (SCCM)

• Microsoft Active Directory, DNS, and DHCP

• vSphere and its components

• Microsoft SQL Server 2008 R2

http://www.vmware.com/�




Installing VMware components

The VMware View Installation Guide available on the VMware website has detailed procedures to install View Connection Server and View Composer 2.5. There are no special configuration instructions required for this solution.

The ESX Installable and vCenter Server Setup Guide available on the VMware website has detailed procedures to install vCenter Server and ESX and is not covered in further detail in this paper. There are no special configuration instructions required for this solution.

Before deploying the desktop pools, ensure that the following steps from the VMware View Installation Guide have been completed:

• Prepare Active Directory

• Install View Composer 2.5 on vCenter Server

• Install View Connection Server (standard and replica)

• Add a vCenter Server instance to View Manager

One desktop pool is created for each vSphere cluster. Two pools will host 500 desktops and the other will host 1,000 desktops. In this solution, persistent automated desktop pools are used as shown in Figure 19.

Figure 19. Persistent automated desktop pools

To create a persistent automated desktop pool as configured for this solution, complete the following steps:

1. Log in to the VMware View Administration page, which is located at https://server/admin, where “server” is the IP address or DNS name of the View Manager server.

2. Click the Pools link in the left pane.

3. Click Add under the Pools banner.

4. In the Type page, select Automated Pool as shown in Figure 20, and click Next.

VMware View installation overview

VMware View setup

VMware View desktop pool configuration

https://server/admin�




50

Figure 20. Select Automated Pool

5. In the User assignment page, select Dedicated and select the Enable automatic assignment checkbox as shown in Figure 21, and click Next.

Figure 21. User assignment




6. In the vCenter Server page, select View Composer linked clones and select a vCenter Server that supports View Composer, as shown in Figure 22. Click Next.

Figure 22. Select View Composer linked clones

7. In the Pool Identification page, enter the required information as shown in Figure 23, and click Next. The pool ID is used by the View Administrators and the Display name is what the users will see in the View Client.




52

Figure 23. Pool identification

8. In the Pool Settings page, make any required changes as shown in Figure 24, and click Next.

Figure 24. Pool settings

9. In the View Composer Disks page, select Do not redirect Windows profile, and click Next.




Figure 25. Select Do not redirect Windows profile

10. In the Provisioning Settings page, select a name for the desktop pool and enter the number of desktops to provision, as shown in Figure 26. Click Next. {n:fixed=4} increments the desktop numbering with 4 digits padded. We used the pool ID at the end to easily associate the desktop name to its pool.

Figure 26. Provisioning settings




54

11. In the vCenter Settings page, browse to select a default image, a folder for the virtual machines, the cluster hosting the virtual desktops, the resource pool to hold the desktops, and the data stores that will be used to deploy the desktops as shown in Figure 27, and then click Next.

Figure 27. vCenter settings

12. In the Select Datastores page, select the datastores for linked clone images, and then click OK. We used Aggressive as the Storage Overcommit option to allow more desktops per virtual provisioned datastore as shown in Figure 28.




Figure 28. Select the datastores for linked clone images

13. In the Guest Customization page, select the domain and AD container, and then select Use a customization specification (Sysprep). Click Next.

Figure 29. Guest customization

14. In the Ready to Complete page (shown in Figure 30), verify the settings for the pool, and then click Finish to start the deployment of the virtual desktops.




56

Figure 30. Verify your settings

PowerPath/VE 5.4.1 supports ESX 4.1. The EMC PowerPath/VE for VMware vSphere Installation and Administration Guide available on Powerlink® provides the procedure to install and configure PowerPath/VE. There are no special configuration instructions required for this solution.

The PowerPath/VE binaries and support documentation are available on Powerlink. Figure 31 shows that PowerPath is managing the block devices on the ESX host.

Figure 31. PowerPath as the owner for managing the path of block devices

PowerPath Virtual Edition




Installing storage components

Storage pools in the EMC VNX OE support heterogeneous drive pools. In this solution, we configured a 96-disk pool with RAID 10 from 92 SAS disks and four near-line SAS drives. From this storage pool, we created 8 thin LUNs, each 2,047 GB in size as shown in Figure 32.

Figure 32. Thin LUNs created

Create storage pools




58

For each LUN in the storage pool, the tiering policy is set to Auto Tiering as shown in Figure 33. As data ages and is used infrequently, it is moved to the near-line SAS drives in the pool.

Figure 33. Auto-Tiering




FAST Cache is enabled as an array-wide feature in the system properties of the array in Unisphere as shown in Figure 34.

Figure 34. Enabling FAST Cache

Enable FAST Cache




60

From the Storage System Properties dialog box, click the FAST Cache tab, click Create, and then select the eligible Flash drives to create the FAST Cache as shown in Figure 35. There are no user-configurable parameters for the FAST Cache.

Figure 35. FAST Cache configuration

FAST Cache is enabled for all LUNs in this solution. The replica images are provisioned on all datastores allocated to that pool. But, as the data gets frequently accessed, it ends up in FAST Cache.




To configure the FAST VP feature for a pool LUN, go to the properties for a pool LUN in Unisphere, click the Tiering tab, and set the tiering policy for the LUN as shown in Figure 36.

Figure 36. Configuring FAST VP

Configure FAST VP




62

The VNX Home Directory installer is available on the NAS Tools website and the application CD for each VNX OE for file release. You can also download the software from Powerlink.

With this feature, you can create a unique share called “HOME,” redirect data to this path based on specific criteria, and provide the user with exclusive rights to the folder.

After installing the VNX Home Directory feature, use the Microsoft Management Console (MMC) snap-in to configure the feature. Figure 37 shows a sample configuration.

Figure 37. Configure the VNX Home Directory feature

The sample configuration shown in Figure 38 shows how to automatically create a user Home Directory for any user in domain view45 in the Homedirs folder on the View45 file system.

\View45\Homedirs\<user>

For example, when user1 logs in, that user will see that \\VNXFILE\HOME points to \View45\Homedirs\User1 on the Data Mover. For user2, \\VNXFILE\Home points to \View45\Homedirs\User2.

Configure VNX Home Directory




Figure 38. Sample Home Directory configuration

Chapter 6: Testing and Validation



64

6 Testing and Validation This chapter compares how the following use cases performed in the boot storm, Login VSI, and antivirus scan test scenarios.

• FAST Cache with no dedicated replica LUN

• FAST Cache and dedicated replica LUN

• No FAST Cache with a dedicated replica LUN

Use case descriptions

In Use Case 1, we created the linked clone desktop pool without a dedicated replica LUN and used 14 flash drives for the FAST Cache configuration. We created a replica virtual machine created for every LUN that hosts the linked clone as shown in Figure 39.

Figure 39. FAST Cache with no dedicated replica LUN

Use Case 1: FAST Cache with no dedicated replica LUN




In Use Case 2, we created one replica virtual machine for every linked clone desktop pool and stored that replica on a different LUN than the linked clones. Figure 40 shows this configuration. We used four Flash drives to host the replica virtual machine and configured FAST Cache to use 10 Flash drives.

Figure 40. FAST Cache with dedicated replica LUNs

Use Case 2: FAST Cache with dedicated replica LUNs




66

In Use Case 3, we did not use FAST Cache and reduced the dedicated replica LUN configuration, hosting 1,000 users in this environment as shown in Figure 41.

Figure 41. Dedicated replica LUN with no FAST Cache

Use Case 3: A dedicated replica LUN with no FAST Cache




Boot storm scenario

This section describes the boot storm results for each of the three use cases when powering up the desktop pools.

For Use Case 1, the virtual desktops took an average of 1.5 seconds to boot. Figure 42 shows the LUN IOPS and response times. The LUN response time stayed below 2 ms.

Figure 42. LUN IOPS and response times for FAST Cache with no dedicated replica LUN

0

2

4

6

8

10

12

14

16

18

20

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

tim

e (m

s)

IOPS

LUN IOPS and response times

ViewDS8 Read IOPS ViewDS8 Write IOPS

ViewDS8 Response time (ms)

Overview





68

Figure 43 shows one of the underlying disk IOPS and response times of the block storage pool.

Figure 43. Physical disk IOPS and response times

Almost 90 percent of the read and write operations are served by the FAST Cache as shown in Figure 44.

Figure 44. FAST Cache read and write operations

0

5

10

15

20

25

30

35

40

0

10

20

30

40

50

60

1 5 9 13172125293337414549535761656973778185899397

resp

onse

tim

e (m

s)

IOPS

Physical disk IOPS and response times

Bus 1 Enclosure 6 Disk 2 - Read Throughput (IO/s)

Bus 1 Enclosure 6 Disk 2 - Write Throughput (IO/s)

Bus 1 Enclosure 6 Disk 2 - Response Time (ms)

0

10

20

30

40

50

60

70

80

90

100

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

VDI - FAST Cache Read Hit Ratio VDI - FAST Cache Write Hit Ratio




Figure 45 shows that the service processor (SP) utilization during the boot storm is approximately 30 percent.

Figure 45. SP utilization during boot storm

An ESX server’s SP utilization remained below 30 percent during the boot process as displayed in Figure 46.

Figure 46. Example boot-time SP utilization

0

5

10

15

20

25

30

35

40

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

SP utilization during boot storm

SP A - Utilization (%) SP B - Utilization (%)

0

5

10

15

20

25

30

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

134

141

148

155

162

169

176

\\c1b1\Physical Cpu(_Total)\% Util Time





70

Figure 47 shows the memory activity from one of the ESX servers. When the virtual machine boots, it consumes the free available memory. The amount of swap memory it uses is very low compared to the memory gain achieved by the Transparent Page Sharing.

Figure 47. ESX memory activity

Figure 48 shows the ESX disk IOPS and response time for one of the LUNs.

Figure 48. ESX physical disk IOPS and guest latency

05000

100001500020000250003000035000400004500050000

1 9 17 25 33 41 49 57 65 73 81 89 97 105

113

121

129

137

145

153

161

169

ESX memory

\\c1b1\Memory\Free MBytes

\\c1b1\Memory\PShare Shared MBytes

\\c1b1\Memory\Swap Used MBytes

\\c1b1\Memory\Total Compressed MBytes

0

5

10

15

20

0

200

400

600

800

1000

1200

1 10 19 28 37 46 55 64 73 82 91 100

109

118

127

136

145

154

163

172

Avg

. Gue

st L

aten

cy M

illis

ec/c

omm

and

IOPS

ESX physical disk IOPS and guest latency

\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\Reads/sec

\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\Writes/sec




Figure 49 shows the number of SCSI reservations prevented by using the Autonomic Test and Set (ATS) feature used by VAAI and the zero requests for the array.

Figure 49. ESX VAAI statistics

This graphic shows that there are approximately 1,670 ATS requests during the boot and about 125 zeroing requests on this LUN.

Figure 50 shows the LUN IOPS and response times during the virtual desktop boot process. The response times stayed below 4 ms most of the time.

Figure 50. LUN IOPS and response time

408204084040860408804090040920409404096040980410004102041040

47000

47500

48000

48500

49000

49500

50000

50500

1 11 21 31 41 51 61 71 81 91 101

111

121

131

141

151

161

171

Zero

s

ATS

ESX VAAI statistics

\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\ATS\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\Zeros

0

2

4

6

8

10

12

14

16

18

20

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55

Resp

onse

Tim

e (m

s)

IOPS

LUN IOPS7 and response times


ViewDS8 Response time

Use Case 2: FAST Cache with dedicated replica LUNs




72

Figure 51 shows the replica LUN IOPS and response times. As expected, most of the IOPS from the replica LUN are “read” IOPS.

Figure 51. Replica LUN IOPS and response times

Figure 52 shows one of the underlying physical disk IOPS and response times of the block storage pool.

Figure 52. Physical disk IOPS and response time

0

5

10

15

20

25

30

0

2000

4000

6000

8000

10000

12000

14000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55

Resp

onse

Tim

e (m

s)

IOPS

Replica LUN IOPS and response times

SSD1 Replica Read IOPS SSD1 Replica Write IOPS

SSD1 Replica Response time

0

2

4

6

8

10

12

14

16

18

20

0

20

40

60

80

100

120

140

160

180

200

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55

resp

onse

tim

e (m

s)

IOPS

Physical disk IOPS and response time







Figure 53 shows the FAST Cache hit ratio of the block storage pool during the virtual machine boot process.

Figure 53. FAST Cache hit ratio

As shown in Figure 54, service processor utilization is approximately 50 percent during the boot storm.

Figure 54. SP utilization during the boot storm

0

20

40

60

80

100

120

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55

FAST Cache hit ratio


0

10

20

30

40

50

60

70

80

90

100

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55

SP utilization during the boot storm





74

Figure 55 shows the CPU utilization of one of the ESX hosts during the boot process.

Figure 55. Example ESX host physical CPU utilization

Figure 56 shows the ESX server’s memory consumption and the memory savings using the transparent page sharing.

Figure 56. ESX server memory

0

10

20

30

40

50

60

70

80

90

100

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57

\\c1b1\Physical Cpu(_Total)\% utilization


0

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000

1 3 5 7 9 111315171921232527293133353739414345474951535557

ESX memory








Figure 57 shows the ESX physical disk IOPS and the average guest latency.

Figure 57. ESX linked clone LUN IOPS and average guest latency

Figure 58 shows that the ESX server used 1,546 ATS operations and 84 “zeroing” requests on the linked clone disk during the boot operation.

Figure 58. Linked clone ESX disk VAAI statistics

02468101214161820

0100200300400500600700800900

1000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58

Avg

. Gue

st L

atec

y (m

s/cm

d)

IOPS

ESX linked clone LUN IOPS and average guest latency



\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\Average Guest MilliSec/Command

23500

23520

23540

23560

23580

23600

23620

23640

25000

25500

26000

26500

27000

27500

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58

Zero

s

ATS

Linked clone ESX disk VAAI statistics





76

Figure 59 shows the replica disk IOPS and average guest latency during the boot operation.

Figure 59. ESX replica disk IOPS and average guest latency

Figure 60 shows that on the replica LUN, ESX used 58 ATS operations and no “zeroing” requests during the virtual machine boot process.

Figure 60. ESX replica LUN VAAI statistics

0

200

400

600

800

0

500

1000

1500

2000

2500

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55

Avg

. Gue

st L

aten

cy (m

sec/

cmd)

IOPS

ESX replica disk IOPS and average guest latency

\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\Reads/sec

\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\Writes/sec

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

450

460

470

480

490

500

510

520

530

540

550

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55

Zero

s

ATS

ESX replica LUN VAAI statistics

\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\ATS\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\Zeros




Without FAST Cache, we need more spindles to host 2,000 users. The existing spindles can support up to 1,000 users, so we performed our testing with 1,000 users for this use case.

Figure 61 shows the linked clone LUN IOPS and response times during the boot process.

Figure 61. Linked clone LUN IOPS and response times

0

2

4

6

8

10

12

14

16

18

20

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70

Resp

onse

Tim

e (m

s)

IOPS

Linked clone LUN IOPS and response times

ViewDS8 Read IOPS ViewDS8 Write IOPS ViewDS8 Response time





78

Figure 62 shows the replica LUN IOPS and response times during the boot process for 1,000 users.


Figure 63 shows the physical disk IOPS and response times during the boot process for 1,000 users.


0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

1 4 7 101316192225283134374043464952555861646770

Resp

onse

Tim

e (m

s)

IOPS


SSD1 Read IOPS SSD1 Write IOPS SSD1 Response time

02468101214161820

0

50

100

150

200

250

1 4 7 101316192225283134374043464952555861646770

Resp

onse

Tim

e (m

s)

IOPS








Figure 64 shows the service processor utilization during the boot process.

Figure 64. Service processor (SP) utilization during boot storm

Figure 65 shows the ESX server physical CPU utilization during the boot process.

Figure 65. ESX server physical CPU utilization

0

10

20

30

40

50

60

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70

SP utilization during boot storm


0

5

10

15

20

25

30

35

40

45

1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031

\\c3b1\Physical Cpu(_Total)\% utilization time





80

Figure 66 shows the ESX memory utilization during the boot process.

Figure 66. ESX memory during boot storm

Figure 67 shows the linked close ESX disk IOPS and average guest latency during the boot process.

Figure 67. Linked clone ESX disk IOPS and average guest latency

0

20000

40000

60000

80000

100000

120000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

ESX memory during boot storm





0

5

10

15

20

0

200

400

600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Avg

. Gue

st L

aten

cy (m

s)

IOPS

Linked clone ESX disk IOPS and average guest latency

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003a25bc67dc25e011)\Reads/sec

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003a25bc67dc25e011)\Writes/sec




Figure 68 shows that the ESX server used 1,480 ATS operations and 2,781 “zeroing” requests on the linked clone disk during the boot operation.

Figure 68. Linked clone ESX disk VAAI

Figure 69 shows the ESX replica disk IOPS and average guest latency.

Figure 69. ESX replica disk IOPS and average guest latency

14000

14500

15000

15500

16000

16500

17000

17500

18000

18500

19000

40000

40500

41000

41500

42000

42500

43000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Zero

s

ATS

Linked clone ESX disk VAAI

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003a25bc67dc25e011)\ATS

0100200300400500600700

0500

100015002000250030003500

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Avg

. Gue

st L

aten

cy (m

s)

IOPS

ESX replica disk IOPS and average guest latency






82

Figure 70 shows that the ESX replica used 125 ATS operations and no “zeroing” requests on the linked clone disk during the boot operation.


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

260026202640266026802700272027402760278028002820

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Zero

s

ATS

ESX replica LUN VAAI statistics





Antivirus scan scenario

We installed the McAfee Enterprise Virus Scan command line utility on all of the virtual desktops in our test environment, and executed the script remotely from a central machine.

Note: Although this is not the preferred way to implement an antivirus scanner in a VDI environment, the purpose of this test is to simulate a traditional customer implementation.

This section describes the antivirus scan results when powering up the desktop pools for each of the three use cases. It includes a results summary graph and graphs showing individual results from scanning 100, 200, 300, 500, and 1,000 desktops in each of the three scenarios.

Summary Figure 71 shows the summary results from an antivirus scan for 500, 300, 200, and 100 desktops for a scenario with FAST Cache and no dedicated replicated LUN.

Figure 71. Antivirus scan summary with FAST Cash and no dedicated replica LUN

The graphs in this section shown the antivirus scan response times for each of the above desktop configurations.

0:00:00

0:07:12

0:14:24

0:21:36

0:28:48

0:36:00

0:43:12

1 23 45 67 89 111

133

155

177

199

221

243

265

287

309

331

353

375

397

419

441

463

485

Tim

e Ta

ken

to S

can

(h:m

m:s

s)

Anti-virus scan summary with FAST Cash and no dedicated replica LUN

500 Scan 300 Scan 200 Scan 100 Scan

Overview





84

100-desktop antivirus scan Figure 72 shows that the 100-desktop scan took 11 minutes.

Figure 72. Scan 100 LUN IOPS and response times

Figure 73 shows the physical disk IOPS and response times for the 100-desktop antivirus scan.


0

2

4

6

8

10

12

14

16

18

20

0

200

400

600

800

1000

1200

1400

1600

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS

Scan 100 LUN IOPS and response times


ViewDS1 Response Time (ms)

02468101214161820

0

1

2

3

4

5

6

7

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS








Figure 74 shows the FAST Cache hit ratio for the 100-desktop antivirus scan.


Figure 75 shows the service processor utilization for the 100-desktop antivirus scan.

Figure 75. 100-desktop antivirus scan SP utilization

75

80

85

90

95

100

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97



0

5

10

15

20

25

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

Scan 100 - SP utilization





86

Figure 76 shows the ESX Processor Utilization for the 100-desktop antivirus scan.

Figure 76. ESX processor utilization

Figure 77 shows the ESX server memory utilization during the 100-desktop antivirus scan.

Figure 77. ESX server memory utilization

0

5

10

15

20

25

30

35

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

106

111

116



0

10000

20000

30000

40000

50000

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

106

111

116

Scan 100 - ESX memory


\\c1b1\Memory\Memctl Current MBytes







Figure 78 shows the ESX LUN IOPS and average guest latency for the 100-desktop antivirus scan.

Figure 78. 100-desktop antivirus scan ESX LUN IOPS and average guest latency

Figure 79 shows the ESX VAAI statistics for the 100-desktop antivirus scan.

Figure 79. Scan 100 - ESX VAAI statistics

0

5

10

15

20

0

500

1000

1500

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

Scan 100 - ESX LUN IOPS and average guest latency



2910029120291402916029180292002922029240292602928029300

4058040600406204064040660406804070040720407404076040780

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

Zero

s

ATS

Scan 100 - ESX VAAI statistics

\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\ATS




88

Figure 80 shows the virtual machine disk IOPS and response times for the 100-desktop antivirus scan.

Figure 80. Scan 100 - virtual machine disk IOPS and response times

200-desktop antivirus scan Figure 81 shows that it took 27 minutes and 12 seconds to scan 200 desktops.

Figure 81. 200-desktop antivirus scan LUN IOPS and response times

0

5

10

15

20

050

100150200250300350400450

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

Read

Res

pons

e Ti

me

(ms)

IOPS

Scan 100 - virtual machine disk IOPS and response times

\\c1b1\Virtual Disk(VD0014P1)\Average MilliSec/Read

\\c1b1\Virtual Disk(VD0014P1)\Reads/sec

\\c1b1\Virtual Disk(VD0014P1)\Writes/sec

0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52

Resp

onse

Tim

e (m

s)

IOPS

Scan 200 - LUN IOPS and response times







Figure 82. 200-desktop antivirus scan physical disk IOPS and response times


Figure 83. 200-desktop antivirus scan FAST Cache Hit ratio

02468101214161820

0

10

20

30

40

50

60

70

80

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52

Resp

onse

Tim

e (m

s)

IOPS

Scan 200 - Physical disk IOPS and response times




70

75

80

85

90

95

100

105

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53

Scan 200 - FAST Cache Hit ratio





90


Figure 84. 200 desktop antivirus scan SP utilization

Figure 85 shows the ESX Server’s processor utilization during the 200-desktop antivirus scan.

Figure 85. ESX server processor utilization

0

10

20

30

40

50

60

70

80

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53



0

10

20

30

40

50

60

70

80

90

100

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53






Figure 86 shows the ESX server’s memory utilization during the 200-desktop antivirus scan.

Figure 86. 200-desktop antivirus scan ESX memory utilization



0

10000

20000

30000

40000

50000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53







0

5

10

15

20

0

500

1000

1500

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS







92


Figure 88. ESX VAAI statistics

Figure 89 shows the virtual machine disk IOPS and latency for the 200-desktop antivirus scan.

Figure 89. 200-desktop antivirus scan virtual machine disk IOPS and latency

29625

29630

29635

29640

29645

29650

29655

41164

41166

41168

41170

41172

41174

41176

41178

41180

41182

41184

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52

Zero

s

ATS



02468101214161820

020406080

100120140160180200

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52

Avg

. Rea

d La

tenc

y

IOPS

Scan 200 - virtual machine disk IOPS and latency










Figure 91. 300-desktop antivirus scan physical disk IOPS and response time

0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77

Resp

onse

Tim

e (m

s)

IOPS

Scan 300 - LUN IOPS and response time



02468101214161820

0

10

20

30

40

50

60

70

80

90

1 4 7 1013161922252831343740434649525558616467707376

Resp

onse

Tim

e (m

s)

IOPS

Scan 300 - physical disk IOPS and response time







94


Figure 92. 300-desktop antivirus scan FAST Cache hit ratio



0

20

40

60

80

100

120

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79

Scan 300 - FAST Cache hit ratio


0

10

20

30

40

50

60

70

80

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79






Figure 94 shows the EXS server CPU utilization for the 300-desktop antivirus scan.

Figure 94. 300-desktop antivirus scan ESX CPU utilization

Figure 95 shows the ESX memory utilization for the 300-desktop antivirus scan.


0

10

20

30

40

50

60

70

80

90

100

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76

Scan 300 - ESX CPU utilization


0

10000

20000

30000

40000

50000

60000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76










96



Figure 97 shows ESX VAAI statistics for the 300-desktop antivirus scan.

Figure 97. 300-desktop antivirus scan

0

5

10

15

20

0

500

1000

1500

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





29600

29620

29640

29660

29680

29700

29720

29740

29760

41170

41175

41180

41185

41190

41195

41200

41205

41210

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77

Zero

s

ATS








500-desktop antivirus scan Figure 99 shows that it took one hour and 17 minutes to scan 500 desktops.


0

2

4

6

8

10

12

14

16

18

20

0

50

100

150

200

250

300

1 4 7 1013161922252831343740434649525558616467707376

Read

Lat

ency

IOPS





0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS







98





02468101214161820

0

20

40

60

80

100

120

140

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS





0

20

40

60

80

100

120

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97







Figure 102. 500-desktop antivirus scan service processor utilization

Figure 103 shows the ESX server CPU utilization during the 500-desktop antivirus scan.

Figure 103. 500-desktop antivirus scan ESX server CPU utilization

0

10

20

30

40

50

60

70

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97



0

10

20

30

40

50

60

70

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

133






100

Figure 104 shows the ESX memory utilization during the 500-desktop antivirus scan.

Figure 104. 500-desktop antivirus scan ESX server memory utilization

Figure 105 shows the ESX disk IOPS and average guest latency for the 500-desktop antivirus scan.

Figure 105. 500-desktop antivirus scan ESX disk IOPS and average guest latency

0

10000

20000

30000

40000

50000

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

133

Scan 500 - ESX memory utilization






05101520253035

0

500

1000

1500

2000

2500

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

134

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

Scan 500 - ESX disk IOPS and average guest latency








Figure 106. 500-desktop antivirus scan ESX VAAI statistics



290002920029400296002980030000302003040030600308003100031200

41160

41180

41200

41220

41240

41260

41280

41300

41320

1 9 17 25 33 41 49 57 65 73 81 89 97 105

113

121

129

137

Zero

s

ATS



0

5

10

15

20

25

30

0

50

100

150

200

250

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

134

Read

Lat

ecy

(ms/

read

)

IOPS








102

Summary Figure 108 shows the summary results from an antivirus scan for 500, 300, 200, and 100 desktops for a scenario with FAST Cache and a dedicated replicated LUN.

Figure 108. Antivirus scan summary with FAST Cache and a dedicated replica LUN

The graphs in this section show the antivirus scan response times for each of the above desktop configurations.

0:00:00

0:14:24

0:28:48

0:43:12

0:57:36

1:12:00

1:26:24

1:40:48

1:55:12

1 23 45 67 89 111

133

155

177

199

221

243

265

287

309

331

353

375

397

419

441

463

485

Tim

e Ta

ken

to S

can

(h:m

m:s

s)

Anti-virus scan summary with FAST Cache and a dedicated replica LUN


Use Case 2: With FAST Cache and a dedicated replica LUN





Figure 109. 100-desktop antivirus scan linked clone LUN IOPS and response times

Figure 110 shows the replica LUN IOPS and response times for the 100-desktop antivirus scan.

Figure 110. 100-desktop antivirus scan replica LUN IOPS and response times

02468101214161820

0

200

400

600

800

1000

1200

1400

1600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Resp

onse

Tim

e (m

s)

IOPS

Scan 100 - Linked clone LUN IOPS and response times



0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Resp

onse

Tim

e (m

s)

IOPS

Scan 100 - Replica LUN IOPS and response times

SSD2 Read IOPS SSD2 Write IOPS SSD2 Response Time (ms)




104





02468101214161820

0

1

2

3

4

5

6

7

8

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Resp

onse

Tim

e (m

s)

IOPS





88

90

92

94

96

98

100

102

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33






Figure 113 shows the service processor utilization for the 100-desktop antivirus scan. The replica LUN is owned by service processor A.




0

5

10

15

20

25

30

35

40

45

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35



0

5

10

15

20

25

30

35

40

45

1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930






106



Figure 116 shows the ESX linked clone LUN IOPS and average guest latency for the 100-desktop antivirus scan.

Figure 116. 100-desktop antivirus scan ESX linked clone LUN IOPS and average guest latency

0

10000

20000

30000

40000

50000

60000

1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930







0

5

10

15

20

0

50

100

150

200

250

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

Scan 100 - ESX linked clone LUN IOPS and average guest latency








Figure 117. 100-desktop antivirus scan ESX VAAI statistics

Figure 118 shows the ESX replica LUN IOPS and average guest latency for the 100-desktop antivirus scan.

Figure 118. 100-desktop antivirus scan ESX replica LUN IOPS and average guest latency

5940

5960

5980

6000

6020

6040

6060

18218

18220

18222

18224

18226

18228

18230

18232

18234

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Zero

s

ATS



0

5

10

15

20

0

500

1000

1500

2000

2500

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

Scan 100 - ESX replica LUN IOPS and average guest latency



\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\Average Guest MilliSec/Command




108

Figure 119 shows the ESX replica LUN VAAI statistics for the 100-desktop antivirus scan.

Figure 119. 100-desktop antivirus scan ESX replica LUN VAAI statistics



0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

50

100

150

200

250

300

350

400

450

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Zero

s

ATS

Scan 100 - ESX replica LUN VAAI statistics


02468101214161820

0200400600800

100012001400160018002000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Resp

onse

Tim

e (m

s)

IOPS











0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Resp

onse

Tim

e (m

s)

IOPS


SSD2 Read IOPS SSD2 Write IOPS SSD2 Response Time

02468101214161820

0

2

4

6

8

10

12

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Resp

onse

Tim

e (m

s)

IOPS








110



Figure 124 shows the service processor utilization during the 200-desktop antivirus scan.


82

84

86

88

90

92

94

96

98

100

102

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64



0

5

10

15

20

25

30

35

40

45

50

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64






Figure 125 shows the ESX CPU utilization during the 200-desktop antivirus scan.




0

5

10

15

20

25

30

35

40

45

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64



0

10000

20000

30000

40000

50000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64










112

Figure 127 shows the ESX server LUN IOPS and average guest latency for the 200-desktop antivirus scan.


Figure 128 shows the ESX linked clone LUN VAAI statistics of the 200-desktop antivirus scan.

Figure 128. 200-desktop antivirus scan ESX linked clone LUN VAAI statistics

0

5

10

15

20

0

100

200

300

400

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





5950

6000

6050

6100

6150

6200

6250

6300

18228

18230

18232

18234

18236

18238

18240

18242

18244

18246

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Zero

s

ATS

Scan 200 - ESX linked clone LUN VAAI statistics





Figure 129 shows the ESX replica LUN IOPS and average guest latency of the 200-desktop antivirus scan.


Figure 130 shows the ESX replica LUN VAAI statistics of the 200-desktop antivirus scan.


0

10

20

30

40

0

1000

2000

3000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





00.10.20.30.40.50.60.70.80.91

0

50

100

150

200

250

300

350

400

450

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64

Zero

s

ATS






114

300-desktop antivirus scan Figure 131 shows that it took 1 hour, 7 minutes, and 39 seconds to scan 300 desktops.


Figure 132 shows the replica LUN IOPS and response time for the 300-desktop antivirus scan.


0

2

4

6

8

10

12

14

16

18

20

0

500

1000

1500

2000

2500

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77

Resp

onse

Tim

e (m

s)

IOPS




0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77

Resp

onse

Tim

e (m

s)

IOPS






Figure 133 shows the physical disk and IOPS response times for the 300-desktop antivirus scan.




02468101214161820

0

2

4

6

8

10

12

14

16

1 4 7 1013161922252831343740434649525558616467707376

Resp

onse

Tim

e (m

s)

IOPS





0

20

40

60

80

100

120

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79






116



Figure 136 shows the ESX CPU utilization during the 300-desktop antivirus scan.


0

5

10

15

20

25

30

35

40

45

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79



0

5

10

15

20

25

30

35

40

45

50

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82








Figure 138 shows the ESX linked close LUN IOPS and average guest latency during the 300-desktop antivirus scan.


0

10000

20000

30000

40000

50000

1 4 7 10131619222528313437404346495255586164677073767982







0

5

10

15

20

0

200

400

600

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS








118

Figure 139 shows the ESX linked clone LUN VAAI statistics for the 300-desktop antivirus scan.


0

1000

2000

3000

4000

5000

6000

7000

8000

9000

18180

18200

18220

18240

18260

18280

18300

18320

18340

18360

18380

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Zero

s

ATS









Figure 141. 300-desktop antivirus scan VAAI statistics for the replica LUN

0

20

40

60

0

1000

2000

3000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

50

100

150

200

250

300

350

400

450

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Zero

s

ATS






120

500-desktop antivirus scan Figure 142 shows that it took 2 hours, 1 minute, and 1 second to scan 500 virtual desktops.




0

2

4

6

8

10

12

14

16

18

20

0

500

1000

1500

2000

2500

3000

3500

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS




0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96

Resp

onse

Tim

e (m

s)

IOPS


SSD2 Read IOPS SSD2 Write IOPS SSD2 Reponse Time (ms)








02468101214161820

02468

101214161820

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS





0

20

40

60

80

100

120

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97






122





0

5

10

15

20

25

30

35

40

45

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97



0

5

10

15

20

25

30

35

40

45

50

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

133








Figure 149 shows the ESX linked clone LUN IOPS and average guest latency during the 500-desktop antivirus scan.


0

10000

20000

30000

40000

50000

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

133







0

5

10

15

20

0

200

400

600

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

134

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS








124



0

2000

4000

6000

8000

10000

12000

183001832018340183601838018400184201844018460184801850018520

1 9 17 25 33 41 49 57 65 73 81 89 97 105

113

121

129

Zero

s

ATS










0

50

100

150

0500

1000150020002500

1 9 17 25 33 41 49 57 65 73 81 89 97 105

113

121

129

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





00.10.20.30.40.50.60.70.80.91

0

50

100

150

200

250

300

350

400

450

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

134

Zero

s

ATS


\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\ATS

\\c1b1\Physical Disk SCSI Device(naa.6006016006202a00a3243c51d12fe011)\Zeros




126

Summary Figure 153 shows the summary results from an antivirus scan for 500, 300, 200, and 100 desktops for a scenario with a dedicated replica LUN but no FAST Cache configured.

Figure 153. Antivirus scan summary without FAST Cash but with a dedicated replica LUN

The graphs in this section show the antivirus scan response times for each of the above desktop configurations.

0:00:00

0:14:24

0:28:48

0:43:12

0:57:36

1:12:00

1:26:24

1:40:48

1:55:12

1 23 45 67 89 111

133

155

177

199

221

243

265

287

309

331

353

375

397

419

441

463

485

Tim

e ta

ken

to s

can

(h:m

m:s

s)

Anti-virus scan summary with a dedicated replica LUN and no FAST Cache






100-desktop antivirus scan Figure 154 shows the LUN IOPS and response times for the 100-desktop antivirus scan.




0

2

4

6

8

10

12

14

16

18

20

0

200

400

600

800

1000

1200

1400

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Resp

onse

Tim

e (m

s)

IOPS




0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Resp

onse

Tim

e (m

s)

IOPS


SSD1 Read IOPS SSD1 Write IOPS SSD1 Response Time




128





02468101214161820

0

10

20

30

40

50

60

70

80

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Resp

onse

Tim

e (m

s)

IOPS





0

5

10

15

20

25

30

35

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35










0

5

10

15

20

25

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35



0

20000

40000

60000

80000

100000

120000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35










130





0

5

10

15

20

0

100

200

300

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS




0

2000

4000

6000

8000

10000

12000

14000

16000

18000

0

5000

10000

15000

20000

25000

30000

35000

40000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Zero

s

ATS


\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003a25bc67dc25e011)\ATS








0

5

10

15

20

0

1000

2000

3000

4000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





00.10.20.30.40.50.60.70.80.91

0

500

1000

1500

2000

2500

3000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Zero

s

ATS


\\c3b1\Physical Disk SCSI Device(naa.6006016006202a00a2243c51d12fe011)\ATS




132

200-descktop antivirus scan Figure 164 shows the LUN IOPS and response times for the 200-desktop antivirus scan.

Figure 164. 200-desktop antivirus scan LUN IOPS and response time



0

2

4

6

8

10

12

14

16

18

20

0

200

400

600

800

1000

1200

1400

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Resp

onse

Tim

e (m

s)

IOPS

Scan 200 - LUN IOPS and response time



0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Resp

onse

Tim

e (m

s)

IOPS







Figure 166. 200-desktop antivirus scan physical disk IOPS and response time



0

2

4

6

8

10

12

14

16

18

20

0

10

20

30

40

50

60

70

80

90

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Resp

onse

Tim

e (m

s)

IOPS

Scan 200 - Physical disk IOPS and response time




0

5

10

15

20

25

30

35

40

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82






134

Figure 168 shows the ESX server’s CPU utilization during the 200-desktop antivirus scan.




0

5

10

15

20

25

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82



0

20000

40000

60000

80000

100000

120000

1 4 7 10131619222528313437404346495255586164677073767982












0

5

10

15

20

0

100

200

300

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS


\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Reads/sec

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Writes/sec

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Average Guest MilliSec/Command




136



0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

0

5000

10000

15000

20000

25000

30000

35000

40000

45000

50000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Zero

s

ATS


\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\ATS\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Zeros





Figure 172. 200-desktop antivirus scanESX replica LUN IOPS and average guest latency



0

5

10

15

20

0

1000

2000

3000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





00.10.20.30.40.50.60.70.80.91

0

500

1000

1500

2000

2500

3000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Zero

s

ATS






138

300-descktop antivirus scan Figure 174 shows that it took 58 minutes and 14 seconds to scan 300 virtual desktops.




02468101214161820

0

200

400

600

800

1000

1200

1400

1 4 7 1013161922252831343740434649525558616467

Resp

onse

Tim

e (m

s)

IOPS




0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1 4 7 1013161922252831343740434649525558616467

Resp

onse

Tim

e (m

s)

IOPS










02468101214161820

0102030405060708090

100

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67

Resp

onse

Tim

e (m

s)

IOPS





0

5

10

15

20

25

30

35

40

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67






140

Figure 178 shows ESX server’s CPU utilization during the 300-desktop antivirus scan.




0

5

10

15

20

25

30

35

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67



0

20000

40000

60000

80000

100000

120000

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67










Figure 180 shows the ESX link clone LUN IOPS and average guest latency for the 300-desktop antivirus scan.




0

5

10

15

20

0

50

100

150

200

250

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS




15400

15420

15440

15460

15480

15500

15520

15540

37989

37989.5

37990

37990.5

37991

37991.5

37992

37992.5

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69

Zero

s

ATS


\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003a25bc67dc25e011)\ATS\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003a25bc67dc25e011)\Zeros




142

Figure 182 shows the replica LUN IOPS and average guest latency for the 300-desktop antivirus scan.

Figure 182. 300-desktop antivirus scan replica LUN IOPS and average guest latency

Figure 183 shows the replica VAAI statistics for the 300-desktop antivirus scan.

Figure 183. 300-desktop antivirus scan replica LUN VAAI statistics

0

20

40

60

80

0

1000

2000

3000

4000

5000

1 4 7 1013161922252831343740434649525558616467

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

Scan 300 - Replica LUN IOPS and average guest latency




0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

500

1000

1500

2000

2500

3000

1 4 7 1013161922252831343740434649525558616467

Zero

s

ATS

Scan 300 - Replica LUN VAAI statistics





500-desktop antivirus scan Figure 184 shows that it took approximately 1 hour and 47 minutes to scan 500 desktops.



Figure 185. 500-desktop antivirus scan replica LUN IOPS and resonse times

02468101214161820

0

500

1000

1500

2000

2500

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96

Resp

onse

Tim

e (m

s)

IOPS




0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96

Resp

onse

Tim

e (m

s)

IOPS






144

Figure 186 shows the physical disk and response times for the 500-desktop antivirus scan.



Figure 187. 500-desktop antivirus scan service process utilization

02468101214161820

0

20

40

60

80

100

120

140

160

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS





0

5

10

15

20

25

30

35

40

45

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97






Figure 188 shows the ESX server’s CPU utilization during the 500-desktop antivirus scan.




0

5

10

15

20

25

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127



0

20000

40000

60000

80000

100000

120000

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

Scan 500 - ESX memory utilization









146


Figure 190. 500-desktop antivirus scan ESX linked clone LUN IOPS and Average guest latency



0

5

10

15

20

0

100

200

300

400

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

Scan 500 - ESX linked clone LUN IOPS and Average guest latency

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Reads/sec

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Writes/sec

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Average Guest MilliSec/Command

02000400060008000100001200014000160001800020000

44038.4

44038.6

44038.8

44039

44039.2

44039.4

44039.6

44039.8

44040

1 9 17 25 33 41 49 57 65 73 81 89 97 105

113

121

129

Zero

s

ATS


\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\ATS

\\c3b1\Physical Disk SCSI Device(naa.6006016007b029003e25bc67dc25e011)\Zeros








020406080100

0

500

1000

1500

2000

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS





00.10.20.30.40.50.60.70.80.91

0

500

1000

1500

2000

2500

3000

1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106

113

120

127

Zero

s

ATS






148

Having the larger FAST Cache configuration benefited Use Case 1 (FAST Cache only) in comparison to Use Cases 2 and 3 as shown in Figure 194.

Figure 194. Average time to scan a single desktop

Note: We performed Use Case 3 (without FAST Cache) on a server that had twice the ESX memory than the other two use cases.


FAST Cache & Replica on EFD Avg (P1) 0:57:04 0:41:01 0:31:30 0:20:36

No FAST Cache Avg (P3/P4) 0:46:33 0:27:46 0:14:47 0:06:11

FAST Cache only Avg (P1) 0:25:10 0:26:30 0:18:03 0:09:43

0:00:00

0:07:12

0:14:24

0:21:36

0:28:48

0:36:00

0:43:12

0:50:24

0:57:36

1:04:48

Tim

e Ta

ken

Average time to scan a single desktop

Antivirus scenario summary




Login VSI test scenario

To simulate a real-world user workload scenario, the Virtual Session Index (VSI) tool version 2.1 was used. The Login VSI workload can be categorized as light, medium, heavy, and custom. A medium workload was selected for this testing and had the following characteristics:

• Simulates normal user behavior and speeds for medium workload

• It uses Microsoft Office applications, Internet Explorer, Adobe Acrobat Reader, and zip files

• The tasks include launching the application, typing, minimizing, maximizing other application, printing, reading PDF, browsing sites that are flash-based.

This section describes the Login VSI tests for each of the three use cases.

We tested this use case in the following conditions:

• With the Auto Tiering option enabled

• With Performance Tiering option enabled

The results of both conditions are shown below.

With the Auto Tiering option enabled Figure 195 shows the Login VSI test results using the Auto Tiering option.

Figure 195. Auto Tiering Login VSI results

Overview

Use Case 1: with FAST Cache and no dedicated replica LUN




150

With the Performance Tiering option enabled Figure 196 shows the Login VSI results using the Performance Tiering option.

Figure 196. Performance Tiering Login VSI results

The following screens capture results from the performance tier output.

Figure 197 shows the LUN IOPS and response times during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 197. LUN IOPS and response times

0

2

4

6

8

10

12

14

16

18

20

0

100

200

300

400

500

600

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS







Figure 198 shows the physical disk IOPS and response times during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 198. Physical disk IOPS and response time

Figure 199 shows the FAST Cache read hit ratio for the during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 199. FAST Cache read hit ratio

02468101214161820

0

5

10

15

20

25

1 5 9 13172125293337414549535761656973778185899397

Resp

onse

Tim

e (m

s)

IOPS

Physical disk IOPS and response time




0

5000

10000

15000

20000

25000

30000

35000

40000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

FAST Cache read hit ratio

VDI - FAST Cache Read Hits/s VDI - FAST Cache Read Misses/s




152

Figure 200 shows the FAST Cache write hit ratio during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 200. FAST Cache write hit ratio

Figure 201 shows the FAST Cache hit ratio for both read and write activity during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.


0

1000

2000

3000

4000

5000

6000

7000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

FAST Cache write hit ratio

VDI - FAST Cache Write Hits/s VDI - FAST Cache Write Misses/s

80

85

90

95

100

105

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97






Figure 202 shows the service processor utilization during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 202. Service processor utilization

Figure 203 shows the ESX CPU utilization during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 203. ESX CPU utilization

0

10

20

30

40

50

60

70

80

90

100

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

SP utilization


0

5

10

15

20

25

30

35

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45

ESX CPU utilization





154

Figure 204 shows the ESX server’s memory utilization during the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 204. ESX memory utilization

Figure 205 shows the ESX disk IOPS and average guest latency for the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 205. ESX disk IOPS and average guest latency

0

10000

20000

30000

40000

50000

60000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45

ESX memory






0

5

10

15

20

0

200

400

600

800

1000

1200

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

ESX disk IOPS and average guest latency







Figure 206 shows the ESX disk VAAI statistics for the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 206. ESX disk VAAI statistics

Figure 207 shows the virtual machine disk IOPS and latency for the Login VSI test with FAST Cache enabled and no dedicated replica LUN.

Figure 207. Virtual machine disk IOPS and latency

36500

37000

37500

38000

38500

39000

39500

40000

40500

46860

46880

46900

46920

46940

46960

46980

47000

47020

47040

47060

1 3 5 7 9 111315171921232527293133353739414345

Zero

s

ATS

ESX disk VAAI statistics


\\c1b1\Physical Disk SCSI Device(naa.6006016007b029003225bc67dc25e011)\Zeros

0

2

4

6

8

10

0

10

20

30

40

50

60

70

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45

Late

ncy

IOPS

Virtual machine disk IOPS and latency




\\c1b1\Virtual Disk(VD0015P1)\Average MilliSec/Write




156

Figure 208 shows the Login VSI test results for Use Case 2.

Figure 208. Login VSI test results

Figure 209 shows the LUN IOPS and response times for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 209. LUN IOPS and response times

0

2

4

6

8

10

12

14

16

18

20

0

1000

2000

3000

4000

5000

6000

1 5 9 131721252933374145495357616569737781858993

Resp

onse

Tim

e (m

s)

IOPS


ViewDS8 Read IOPS ViewDS8 Write IOPS ViewDS8 Response Time

Use Case 2: with FAST Cache and with a dedicated replica LUNs




Figure 210 shows the replica LUN IOPS and response times for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.


Figure 211 shows the FAST Cache read hit ratio for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 211. FAST Cache read hit ratio

0

2

4

6

8

10

12

14

16

18

20

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

1 5 9 131721252933374145495357616569737781858993

Resp

onse

Tim

e (m

s)

IOPS


SSD1 Replica Read IOPS SSD1 Replica Write IOPS

SSD1 Response time

0

5000

10000

15000

20000

25000

30000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

FAST cache read hit ratio

VDI - FAST Cache Read Hits/s VDI - FAST Cache Read Misses/s




158

Figure 212 shows the FAST Cache write hit ratio for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 212. FAST Cache write hit ratio

Figure 213 shows the FAST Cache hit ratio for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.


0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

FAST Cache write hit ratio

VDI - FAST Cache Write Hits/s VDI - FAST Cache Write Misses/s

0

20

40

60

80

100

120

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97






Figure 214 shows the service processor utilization for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 214. Service processor utilization

Figure 215 shows the ESX server’s CPU utilization during the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 215. ESX server CPU utilization

0

10

20

30

40

50

60

70

80

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

SP utilization


0

10

20

30

40

50

60

70

80

90

100

1 13 25 37 49 61 73 85 97 109

121

133

145

157

169

181

193

205

217

229

241

253

265

277

289

301

ESX CPU utilization





160

Figure 216 shows the ESX server’s memory utilization during the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 216. ESX server memory utilization

Figure 217 shows the ESX linked clone LUN IOPS and average guest latency for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 217. ESX linked clone LUN IOPS and average guest latency

0

10000

20000

30000

40000

50000

60000

1 14 27 40 53 66 79 92 105

118

131

144

157

170

183

196

209

222

235

248

261

274

287

300

ESX memory






0

5

10

15

20

0

500

1000

1500

1 16 31 46 61 76 91 106

121

136

151

166

181

196

211

226

241

256

271

286

301

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

ESX linked clone LUN IOPS and average guest latency







Figure 218 shows the ESX lined clone LUN VAAI statistics for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 218. ESX linked clone LUN VAAI statistics

Figure 219 shows the ESX replica LUN IOPS and average guest latency for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.

Figure 219. ESX replica LUN IOPS and average guest latency

0

5000

10000

15000

20000

25000

243502440024450245002455024600246502470024750248002485024900

1 17 33 49 65 81 97 113

129

145

161

177

193

209

225

241

257

273

289

Zero

s

ATS

ESX linked clone LUN VAAI statistics


0

5

10

15

20

0

200

400

600

800

1000

1 16 31 46 61 76 91 106

121

136

151

166

181

196

211

226

241

256

271

286

301

Avg

. Gue

st L

aten

cy (m

s/cm

d)

IOPS

ESX replica LUN IOPS and average guest latency







162

Figure 220 shows the ESX replica LUN VAAI statistics for the Login VSI test with FAST Cache enabled and a dedicated replica LUN.


This section provides Login VSI results from tests with 1,000 users with a dedicated replica LUN with no FAST Cache.

Figure 221 shows the Login VSI test results for Use Case 3.

Figure 221. Login VSI test results

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

100

200

300

400

500

600

1 16 31 46 61 76 91 106

121

136

151

166

181

196

211

226

241

256

271

286

301

Zero

s

ATS

ESX Replica LUN VAAI statistics






Figure 222 shows the linked clone LUN IOPS and response times for the Login VSI test with a dedicated replica LUN and with no FAST Cache.

Figure 222. Linked clone LUN IOPS and response times

Figure 223 shows the replica LUN IOPS and response times for the Login VSI test with a dedicated replica LUN and with no FAST Cache.


0

2

4

6

8

10

12

14

16

18

20

0

1000

2000

3000

4000

5000

6000

1 11 21 31 41 51 61 71 81 91 101

111

121

131

141

151

161

171

181

191

Resp

onse

Tim

e (m

s)

IOPS

Linked clone LUN IOPS and response times



0

2

4

6

8

10

12

14

16

18

20

0

2000

4000

6000

8000

10000

12000

1 11 21 31 41 51 61 71 81 91 101

111

121

131

141

151

161

171

181

191

Resp

onse

Tim

e (m

s)

IOPS






164

Figure 224 shows the physical disk IOPS and response times for the Login VSI test with a dedicated replica LUN and with no FAST Cache.


Figure 225 shows the service processor utilization for the Login VSI test with a dedicated replica LUN and with no FAST Cache.

Figure 225. Service process utilization

02468101214161820

0

50

100

150

200

250

300

1 11 21 31 41 51 61 71 81 91 101

111

121

131

141

151

161

171

181

191

Resp

onse

Tim

e (m

s)

IOPS





0

10

20

30

40

50

60

70

80

90

100

1 9 17 25 33 41 49 57 65 73 81 89 97 105

113

121

129

137

145

153

161

169

177

185

193

SP utilization


Chapter 7: Conclusion



7 Conclusion This section summarizes the solution test results and includes the following sections:

• Summary

• Findings

• References

Summary EMC’s VNX platform along with VMware View enables customers to host virtual desktops economically and to minimize the risk of exposure of data. The presented solution highlights the design guidelines for hosting 2,000 users on EMC VNX5700 and uses advanced technologies like EMC FAST VP and EMC FAST Cache to optimize the performance for the virtual desktop environment.

Findings The EMC solution team confirmed the following key results during the testing of this solution:

• By using FAST Cache and VAAI, the time to concurrently boot all 2,000 desktops to a usable start is reduced by 25 percent.

• Having the replica datastore use FAST Cache reduces the virus scanning per desktop by almost 50 percent.

• With no dedicated replica LUN and using FAST Cache, the maximum response time of the simulated workload is lower compared to the other two use cases.

• By using a VAAI-enabled storage platform, we are able to store up to 512 virtual machines compared to 64 virtual machines per LUN without VAAI-enabled storage.

• Using Flash as FAST Cache for the read and write I/O operations reduces the number of spindles needed to support the required IOPS.

Chapter 7: Conclusion



166

References Refer to the following white papers, available on Powerlink, for information about solutions similar to the one described in this paper:

• EMC Infrastructure for Virtual Desktops Enabled by EMC VNX Series, VMware vSphere 4.1, VMware View 4.5, and VMware View Composer 2.5—Reference Architecture

• EMC Infrastructure for Virtual Desktops Enabled by EMC VNX Series, VMware vSphere 4.1, VMware View 4.5, and VMware View Composer 2.5—Proven Solution Guide

• Deploying Microsoft Windows 7 Virtual Desktops with VMware View—Applied Best Practices Guide

• EMC Performance Optimization for Microsoft Windows XP for the Virtual Desktop Infrastructure—Applied Best Practices

If you do not have access to the above content, contact your EMC representative.

The following Cisco documents, located on the Cisco website, also provide useful information:

• Cisco Desktop Virtualization Solution Whitepaper (link to http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns836/ns978/solution_overview_c22-632364.pdf )

• Cisco Validated Design for Desktop Virtualization with VMware View and EMC Storage (link to http://www.cisco.com/en/US/solutions/ns340/ns414/ns742/ns743/ns993/landing_dcVirt-VM_EMC.html)

• Cisco Desktop Virtualization Solutions Website (link to www.cisco.com/go/vdi)

• Cisco Virtualization Experience Infrastructure Website (link to www.cisco.com/go/vxi)

The following documents are available on the VMware website:

• Introduction to VMware View Manager

• VMware View Manager Administrator Guide

• VMware View Architecture Planning Guide

• VMware View Installation Guide

• VMware View Integration Guide

• VMware View Reference Architecture

• Storage Deployment Guide for VMware View

• VMware View Windows XP Deployment Guide

• VMware View Guide to Profile Virtualization

White papers

Cisco documentation

Other documentation

http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns836/ns978/solution_overview_c22-632364.pdf�

http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns836/ns978/solution_overview_c22-632364.pdf�

http://www.cisco.com/en/US/solutions/ns340/ns414/ns742/ns743/ns993/landing_dcVirt-VM_EMC.html�

http://www.cisco.com/en/US/solutions/ns340/ns414/ns742/ns743/ns993/landing_dcVirt-VM_EMC.html�

http://www.cisco.com/go/vdi�

http://www.cisco.com/go/vxi�

Download - VMware View, EMC VNX, and Cisco UCS Proven Solution · PDF fileCISCO UNIFIED COMPUTING SYSTEMS . Proven Solution Guide . EMC Infrastructure for Virtual Desktops enabled by EMC VNX,

Top Related