cloud operating system unit 15 cloud resource management m. c. chiang department of computer science...

Cloud Operating System

Unit 15Cloud Resource Management

M. C. Chiang

Department of Computer Science and Engineering National Sun Yat-sen University

Kaohsiung, Taiwan, ROC

Cloud Operating System

Outline What are resources? Resources

Requesting/Managing SOAP REST

Load Balancing Auto-scaling

Shared-Nothing Architecture

Scheduling Crosses Nodes

Network Management

04/19/23 Cloud Operating System - Unit 15: Cloud Resource Management 1 U15-2

What Are Resources

Everything in a host CPU Memory Disk storage Network Bandwidth

Resources should be distributed to instances/users not only fairly, but also properly


Resources Requesting/Managing – URI Uniform Resource Identifier

URL is a type of URI, used for representing resources on internet.

URI can represent more types of resource, including local resources.

Need a standard for parsing the data in a page/resource


Resources Requesting/Managing – Data For web services, standardizing information

exchanging helps simplifying communications between services.

In 1998, Microsoft and UserLand Software created the XML-RPC protocol, and supported by other companies.

SOAP is based on XML-RPC, with more flexibility, and become the part of W3C standard in 2000

JSON is another protocol describing data by using JavaScript


Resources Requesting/Managing – REST REpresentational State Transfer It’s an architectural style, not a protocol Essence of the way the Web already works

For example:

http://example.com/resources represents a collection of resources. And a HTTP GET operation means “list” the resources.

http://example.com/resources/item17 represent a resource in collection. HTTP GET means “retrieve”

the resource.


Resources Requesting/Managing – RESTResource GET PUT POST DELETE

Collection URI, such ashttp://

example.com/resources/

List the URIs and perhaps other details of the collection's members.

Replace the entire collection with another collection.

Create a new entry in the collection. The new entry's URL is assigned automatically and is usually returned by the operation.

Delete the entire collection.

Element URI, such ashttp://

example.com/resources/item17

Retrieve a representation of the addressed member of the collection, expressed in an appropriate Internet media type.

Replace the addressed member of the collection, or if it doesn't exist, create it.

Treat the addressed member as a collection in its own right and create a new entry in it.

Delete the addressed member of the collection.


Resources Requesting/Managing – REST OpenStack, AWS all support RESTful API Example for OpenStack

POST http://localhost/v1/AUTH_admin/mycontainer Create a container “mycontainer”

PUT file1 http://localhost/v1/AUTH_admin/mycontainer/ Upload a file “file1” to “mycontainer”

GET http://localhost/v1/AUTH_admin/mycontainer/?prefix=fi List files in “mycontainer” with prefix fi

GET http://localhost/v1/AUTH_admin/mycontainer/file1 Download the file “file1”

The RESTful way makes better and scalable interface for a system


Resources Requesting/Managing - SOAP Simple Object Access Protocol

It’s a protocol, not a style like REST

No new inventions involved Use XML for describing the content Can be used over any transport protocol such as

HTTP, SMTP, or raw TCP


Resources Requesting/Managing - SOAP A SOAP message

components: SOAP envelope

The wrapper

SOAP body Application-specific

message. The client/server can

interpret it by their specific way


Resources Requesting/Managing - SOAP


A price request from client:1. <SOAP-ENV:Envelope

2. xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"

3. SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">

4. <SOAP-ENV:Header>

5. </SOAP-ENV:Header>

6. <SOAP-ENV:Body>

7. <m:GetLastTradePrice xmlns:m="http://foo.bar/prices">

8. <m:Item>Apple</m:Item>

9. </m:GetLastTradePrice>

10. </SOAP-ENV:Body>

11. </SOAP-ENV:Envelope>

encodingStyle in line 3 is the attribute of envelop describes how the message encoded

Line 5 to 7 is the message wraps the application-specific information It’s a call to “GetLastTradePrice” request for server in line 5

Resources Requesting/Managing - SOAP Response from server:1. <SOAP-ENV:Envelope

2. xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"

3. SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">

4. <SOAP-ENV:Body>

5. <m:GetLastTradePriceResponse xmlns:m=“http://foo.bar/prices">

6. <m:Price>34.5</m:Price>

7. </m:GetLastTradePriceResponse>

8. </SOAP-ENV:Body>

9. </SOAP-ENV:Envelope>

As we can see in Line 5 it’s a “GetLastTradePriceResponse” And in Line 6 the “Price” is 34.5


Resources Requesting/Managing - SOAP A run-instance SOAP request body example in AWS EC2:1. <RunInstances xmlns="http://ec2.amazonaws.com/doc/2012-03-01/">

2. <instancesSet>

3. <item>

4. <imageId>ami-60a54009</imageId>

5. <minCount>1</minCount>

6. <maxCount>3</maxCount>

7. </item>

8. </instancesSet>

9. <groupSet/>

10. </RunInstances>

1. AWS EC2 service will parse the request, and get the variables respective, and execute the RunInstances request


Resource Requesting/Managing With the RESTful API and SOAP, we can

manipulate the resource on the Cloud Computing through Web Service.

After all, Web Service is the final product of Cloud Computing for the end-users.


Load Balancing

Load balancing is very important in resource management

Another “As a Service” in many cloud provider AWS Elastic Load Balancing (ELB) Rackspace Cloud Load Balancing Project Atlas of OpenStack as a pluggable module


Load Balancing

Load balancing balances not only network bandwidth, but also system loading

Also monitoring the health of instances If a instance is unhealthy, the balancer should stop

sharing the load with the instance.

Session Persistence Connections should be bound to the same node in a

connection session For some local data and performance


Auto-scaling The most important feature of Cloud Computing

On-demand self-service

Work with load balancer Must get loading information to determine when to scale.

Easy with virtualization You can’t install a new physical machine automatically.

Scale up for more computing power, and scale down for saving money


Auto-scaling – Issues

Synchronizing between nodes Also an issue in load-balancing

Scales too large Needs to set the budget constraints

That is “Measured service”

Hard to scale down for database With shared-nothing architecture, it’s hard to join the

data sets from different hosts back when shutting down nodes.


Shared-nothing Architecture

Sharing no memories and disks between nodes Each node is self-sufficient Independency

No “Single Point of Failure” What will happen if an NFS server shared by 10 hosts for

storage, and the server crashes?

More easily for scaling than shared-disc/memory


SN Architecture – Database Partitioning the table

with row, rather than column

Distributes the rows to nodes Then the queries can be

distributed

Google calls it “database shard”


id data

1 data1

2 data2

3 data3

4 data4

id data

1 data1

2 data2

id data

3 data3

4 data4

Shared-nothing vs. Shared resourcesShared-Nothing: Better availability Better scalability Easier to replace the fault

device, but complicates maintenance

Need redundant resources

Shared resources: Single Point of

Contention problem Larger resource pool Centralize management,

but leads to longer downtime in maintenance

Complex synchronization and locking design


Scheduling Crosses Nodes In a normal OS, scheduler schedules processes

when to run In Cloud, when incoming a virtual machine

instance request, scheduler is responsible for scheduling which host in which zone the instance should run on

The algorithm can be very simple or very complex

It’s host OS’s responsibility to schedules when VMs running on the host


Scheduling Crosses Nodes – Example OpenStack nova-scheduler algorithms

Chance The host is chosen randomly

Availability zone The host is chosen randomly from within a specified

availability zone

Simple The host whose load is least is chosen. Need load

balancer


Scheduling Crosses Nodes – Example OpenStack nova-scheduler architecture

Using a simple table to maintain instance types (how many memories, etc.) in every host, and use this table to determine how many resources are consumed. It ignores the status updates.

Only one scheduler workers – it leads to “Single Point of Failure”

“Current workload” as a parameter to avoid hammering a heavy-loading host


Scheduling Crosses Nodes

Scheduling is a hot topic in Cloud Computing In a very large implement (500 hosts with 200

instances each, for example), it will be very critical to reduce the scheduling latency

In OpenStack the schedule algorithm is very simple, but has some architectural problems. Better model is released in newest implementation (April 5)


Network Management Network is fundamental of accessing Cloud There is no physical interface for virtual machines.

Only vNICs, then bridge to the real NICs In this model, Linux bridge control only provides

basic management. No advanced QoS, ACLs, and monitoring

Simple and flat network topology. May not suit all productions

In OpenStack, the nova-network leads to the Single Point of Failure problem again


Network Management – Virtual Switch Improved network virtualization

Not only virtualize NICs, but also switches now Virtual switch as a service layer for management Virtual switch provides virtual ports for connectivity

Virtual Switch provides the function normal switch should have Layer 2 management QoS, port statistics Enable more complex network topology


Network Management – Virtual Switch Available implementations

Open vSwitch


Network Management - Example OpenStack Quantum project

Included in Essex release Including APIs, publicly available virtual switch

implements and other plugins Support RESTful API

quantum.foo.com/tenant-a/network/net-1/port/17 We can get statistic of port 17 in net-1 of tenant-a

Through Quantum the network service can be more easily to be measured Network-as-a-Service


Summary

SOAP provides a standard method for exchanging data

REST architecture describes a more general method to manage resources

Load balancing not only can balance the resource, but also can be used for failover

Virtualization makes auto-scaling easy and possible


Summary

Shared-nothing is mainstream in Cloud, since it has higher availability and scalability

Scheduling in Cloud schedules on which host the new virtual machine should be created

As process scheduling is latency-critical in OS, VM scheduling is latency-critical in Cloud as well

Improved network virtualization will be next step in Cloud, enables more capabilities


cloud operating system unit 15 cloud resource management m. c. chiang department of computer science...

Documents

collection of resources

cloud resource management

roc cloud

local resources

entire collection

types of resource

file1 http

openstack post http