distributed storage networks and computer...
TRANSCRIPT
University of FreiburgTechnical FacultyComputer Networks and TelematicsWinter Semester 2011/12
Distributed Storage Networksand Computer Forensics1. Organization & OverviewChristian SchindelhauerAmir Alsbih
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Organization
‣ Lecture • Monday 12:15-13:00, 101/SR 00-010/14• Thursday, 11:15 - 13:00, 101/SR 00-010/14
‣ Exercise (Christian Ortolf)• starts Oct 31, 2011• Monday 13:00-14:00, 101/SR 00-010/14• appear Tuesday on the web-pages• are the bases for the oral exam• solutions of the exercises are discussed in the following week
2
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Web
‣ Web page• http://cone.informatik.uni-freiburg.de/cone_teach/
cone_teach_current/dsacfws11• Slides, exercises, videos, link to forum
‣ Forum• for discussion, links, funnies etc.• http://archive.cone.informatik.uni-freiburg.de/forum3/
viewforum.php?f=6
3
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Exam
‣ Oral exam• based on the lecture and the exercises• closed book exam• selected exercises solutions may be used
‣ Mandatory registration using HIS• Questions in the exam from the lecture and the
exercises
4
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Overview
‣ Basic Storage Technology• Hard disks• Flash memory, solid state disks• Storage device design
‣ File systems• Classic file systems• Network and distributed file
systems‣ Storage organization
• SAN, NAS, FAN• Storage hierarchies, Tiers
‣ Redundancy• RAID levels
• Coding techniques‣ Distributed Storage
• Peer-to-peer network storage- e.g. Oceanstore
‣ Computer Forensics- Foundations- Methods
• Collecting evidence• Windows forensics• Linux forensics
5
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
MotivationEvolution of Disks
Algorithms and Methods for Distributed Storage Networks
6
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Storage Prices 2006
7
Flash vs. hard drives: The battle intensifiesTom Coughlin, Jim Handy, Solid State Technology 2010
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Price Fall of RAM and Disk Storage
8
Flash vs. hard drives: The battle intensifiesTom Coughlin,Jim Handy, Solid State Technology 2010
observed since the first disk drive. Although the CGRis expected to eventually decline, based on increasedprocessing difficulties in magnetic head and disk me-dia production, laboratory demonstrations of ad-vanced head designs indicate that a head with a ca-pacity of greater than 100 gigabits/in2 is feasiblewithin the near future, representing an increase oftwo or three times, compared with the areal densityof today’s production disk drives.3,4
Although there is no direct analogue to areal den-sity at the storage system level, the trend in floorspace utilization in terabytes/ft2 closely approximatesthat of areal density at the drive level. Figure 2 il-lustrates this trend for storage systems and nearlyparallels the increases indicated by Figure 1 in diskdrive areal density. The enhanced storage capacityper square foot of floor space is the direct result ofdisk drive increases in areal density and, as this trendcontinues into the future, it is expected that nearly104 gigabytes/ft2 will be attained before the year2010.
An alternate method of considering this trend isshown in Figure 3, where the floor space required
to store one terabyte of information over the past45 years is shown to have decreased by more thana factor of 107, similar to increases in server disk driveareal density.
As a further example of the relationship between sys-tems containing multiple disk drives and the evolu-tionary trends in the drives themselves, the volumet-ric density of disk drives is shown in Figure 4. Thisparameter combines disk areal density with the pack-ing efficiency of disks within the drive’s frame as wellas the packaging of the spindle motor, actuator mo-tor, and electronics. Normally, drives with a smallerform factor (FF) exhibit the highest volumetric den-sity, due to more efficient packing. A marked changein the slope of the volumetric density curve indicatesthis effect in larger-form-factor (14-inch and 10.8-inch) drives to 3.5-inch and smaller drives. Areal den-sity also influences this slope change.
At the system level, packing large numbers of drivesin close proximity combined with high volumetricdensity of the drives themselves results in the trendshown in Figure 5. The use of smaller-form-factor
Figure 1 Hard disk drive areal density trend
AFC=ANTIFERROMAGNETICALLY COUPLEDGMR=GIANT MAGNETORESISTIVE
RANGE OFPOSSIBLEVALUES
AR
EAL
DEN
SIT
Y M
EGAB
ITS/
SQ
UA
RE
INC
H
105
104
103
102
10
1
10–1
10–2
10–3
106
19701960 1980 1990 2000 2010PRODUCTION YEAR
IBM RAMAC™ (FIRST HARD DISK DRIVE)
~35 MILLION XINCREASE
1st THIN FILM HEAD
25% CGR
60% CGR
100% CGR
1st MR HEAD
1st GMR HEAD
1st AFC MEDIA
IBM DISK DRIVE PRODUCTSINDUSTRY LAB DEMOS
Figure 2 Storage floor space utilization trend — IBM storage systems
RAW
CAP
ACIT
Y P
ER F
LOO
R S
PAC
E A
REA
, GB
YTES
/SQ
UA
RE
FOO
T
103
102
101
100
10–1
10–2
10–3
10–4
104
19701960 1980 1990 2000 2010PRODUCTION YEAR
IBM RAMAC (FIRST HARD DISK DRIVE)
ESS MOD F
3380
RAMAC 2
IBM ESS MOD 800
25% CGR
60% CGR
RANGE OFPOSSIBLE VALUES
ESS=IBM TOTALSTORAGE™ ENTERPRISE STORAGE SERVER™
IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003 GROCHOWSKI AND HALEM 339
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Increase of Density
9
Technological impact of magnetic hard disk drives on storage systems, Grochowski, R. D. Halem IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003
Flash vs. hard drives: The battle intensifiesTom Coughlin,Jim Handy, Solid State Technology 2010
observed since the first disk drive. Although the CGRis expected to eventually decline, based on increasedprocessing difficulties in magnetic head and disk me-dia production, laboratory demonstrations of ad-vanced head designs indicate that a head with a ca-pacity of greater than 100 gigabits/in2 is feasiblewithin the near future, representing an increase oftwo or three times, compared with the areal densityof today’s production disk drives.3,4
Although there is no direct analogue to areal den-sity at the storage system level, the trend in floorspace utilization in terabytes/ft2 closely approximatesthat of areal density at the drive level. Figure 2 il-lustrates this trend for storage systems and nearlyparallels the increases indicated by Figure 1 in diskdrive areal density. The enhanced storage capacityper square foot of floor space is the direct result ofdisk drive increases in areal density and, as this trendcontinues into the future, it is expected that nearly104 gigabytes/ft2 will be attained before the year2010.
An alternate method of considering this trend isshown in Figure 3, where the floor space required
to store one terabyte of information over the past45 years is shown to have decreased by more thana factor of 107, similar to increases in server disk driveareal density.
As a further example of the relationship between sys-tems containing multiple disk drives and the evolu-tionary trends in the drives themselves, the volumet-ric density of disk drives is shown in Figure 4. Thisparameter combines disk areal density with the pack-ing efficiency of disks within the drive’s frame as wellas the packaging of the spindle motor, actuator mo-tor, and electronics. Normally, drives with a smallerform factor (FF) exhibit the highest volumetric den-sity, due to more efficient packing. A marked changein the slope of the volumetric density curve indicatesthis effect in larger-form-factor (14-inch and 10.8-inch) drives to 3.5-inch and smaller drives. Areal den-sity also influences this slope change.
At the system level, packing large numbers of drivesin close proximity combined with high volumetricdensity of the drives themselves results in the trendshown in Figure 5. The use of smaller-form-factor
Figure 1 Hard disk drive areal density trend
AFC=ANTIFERROMAGNETICALLY COUPLEDGMR=GIANT MAGNETORESISTIVE
RANGE OFPOSSIBLEVALUES
AR
EAL
DEN
SIT
Y M
EGAB
ITS/
SQ
UA
RE
INC
H
105
104
103
102
10
1
10–1
10–2
10–3
106
19701960 1980 1990 2000 2010PRODUCTION YEAR
IBM RAMAC™ (FIRST HARD DISK DRIVE)
~35 MILLION XINCREASE
1st THIN FILM HEAD
25% CGR
60% CGR
100% CGR
1st MR HEAD
1st GMR HEAD
1st AFC MEDIA
IBM DISK DRIVE PRODUCTSINDUSTRY LAB DEMOS
Figure 2 Storage floor space utilization trend — IBM storage systems
RAW
CAP
ACIT
Y P
ER F
LOO
R S
PAC
E A
REA
, GB
YTES
/SQ
UA
RE
FOO
T
103
102
101
100
10–1
10–2
10–3
10–4
104
19701960 1980 1990 2000 2010PRODUCTION YEAR
IBM RAMAC (FIRST HARD DISK DRIVE)
ESS MOD F
3380
RAMAC 2
IBM ESS MOD 800
25% CGR
60% CGR
RANGE OFPOSSIBLE VALUES
ESS=IBM TOTALSTORAGE™ ENTERPRISE STORAGE SERVER™
IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003 GROCHOWSKI AND HALEM 339
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Increase of Density
(Floor Space)
10
Technological impact of magnetic hard disk drives on storage systems, Grochowski, R. D. Halem IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Evolution of Disk Form
Factors
11
Technological impact of magnetic hard disk drives on storage systems, Grochowski, R. D. Halem IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003
form-factor drives, such as 2.5-inch drives, will con-stitute the storage devices used for large storage ar-rays.4 In the past, only eight large-form-factor diskdrives could be contained within a storage systemframe of reasonable size, whereas today’s systemsmay contain as many as 256 drives (or more) in aRAID (redundant array of independent disks) orequivalent configuration. The availability of small-form-factor drives has been directly responsible forthe development and use of system architecturessuch as RAID, which have advanced storage to newlevels of low cost, capacity, performance, and reli-ability. In addition to the increases in reliabilitybrought about by these enhanced drive technologies,the adoption of RAID architectures has resulted ineven higher levels of improvement in system avail-ability. RAID architectures enable the effective un-coupling of hardware reliability from system avail-ability.
The miniaturization trend in disk drives, simulta-neous with the trend of increasing drive capacity, isthe direct result of the vast technology improvementsthat have increased areal density over seven ordersof magnitude since 1956. These improvements haveprincipally been: the introduction of magnetoresis-tive (MR) and giant magnetoresistive (GMR) readheads, finer-line-width inductive write head ele-ments, high-signal-amplitude thin film disk materi-als, lower flying heights, and partial-response-max-imum-likelihood (PRML) data channels. These andmany more innovations demonstrate the continuingtrend toward further enhancements.5
Power/performance
Smaller-form-factor drives containing smaller-diam-eter disks with high areal density permit faster diskrotation rates while maintaining moderate power re-quirements. Figure 9 shows this trend for disk drivesand storage systems, and indicates that both havebecome significantly more economical with respectto power utilization. The power loss due to air shearfor a disk drive is given by
P ! constant " D 4.6 " R 2.8
where D is disk diameter and R is the rotation rate.Reducing disk diameter from 14 inches (355 mm)to 65 mm would allow the drive designer to increaserotation rate from 3600 RPM (revolution per minute)to 15000 RPM in today’s higher performance diskdrives, while maintaining acceptable power losses
Figure 8 The evolution of disk drive form factors 1956–2002
24-INCH FORM FACTOR5 MB(NOT SHOWN)
10.8-INCH FORM FACTOR 5.7 GB
14-INCH FORM FACTOR 3.8 GB
1.0-INCH FORM FACTOR 1 GB
8-INCH FORM FACTOR 65 MB
2.5-INCH FORM FACTOR 60 GB
3.25-INCH FORM FACTOR84-MM DISKS/10 000 RPM 146 GB70-MM DISKS/15 000 RPM36 GB
3.25-INCH FORM FACTOR 120 GB
GROCHOWSKI AND HALEM IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003342
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Increase of Speed
12
Technological impact of magnetic hard disk drives on storage systems, Grochowski, R. D. Halem IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003
Overall system performance! !cache hit ratio " channel performance)
# !1 $ cache hit ratio" " drive performance
System performance is a weighted average of the twoperformances based on the cache hit ratio. Nearlyconcurrent with this trend, a disk drive buffer wasalso included, which enhanced the performance ofthis component significantly. Figure 12 shows rela-tive performance of both drives and systems overtime, and the effect of caching and buffering is clearlyvisible. The basis for the chart is the storing of 4Krecords. Although system (and drive) performancewould be expected to increase as a result of com-ponent technology advances and drive miniaturiza-tion, enhancements in electronics have been a sig-nificant factor.
The CGR slopes of the relative performance(input/output operations per second) curves in Fig-ure 12 after 1990 are about 25 percent, less than therate of disk capacity increase based on areal den-sity. There is a hypothetical concern that perfor-mance per capacity could decrease to a level whereoverall system performance would be adversely af-fected. Presently, this performance trend constitutesessentially no limitation to future disk drive capac-ity increase, although new design modifications toenhance disk drive and system performance could
Figure 12 Performance trend for hard disk drives and storage systems
RE
LATI
VE
I/O
’S P
ER
SE
CO
ND
100
10
1
0.1
1000
1970 1980 1990 2000AVAILABILITY YEAR
BASIS: 4K RECORDS
ADDITION OF HDD BUFFER
ADDITION OF SYSTEM CACHE
HARD DISK DRIVE MODELED DATA
STORAGE SYSTEMS
Figure 11 Disk drive access/seek times
TIM
E, M
ILLI
SEC
ON
DS
10
1
100
1970 1975 1980 1985 1990 1995 2000 2005AVAILABILITY YEAR
SEEKINGACCESSING
ULTRASTAR XP
ULTRASTAR 2XP
ULTRASTAR 18XP
ULTRASTAR 9ZX
ULTRASTAR 36XP
ULTRASTAR 18ZX
ULTRASTAR 36ZX
ULTRASTAR 73LZX
ULTRASTAR 36LZX
ULTRASTAR 36Z15
SEEK TIME ~ (ACTUATOR INERTIALPOWER)1/3 x (DATA BAND)2/3
ROTATIONAL TIME (LATENCY) ~ (RPM)-1
ACCESS TIME = SEEK TIME + LATENCY
~
~
GROCHOWSKI AND HALEM IBM SYSTEMS JOURNAL, VOL 42, NO 2, 2003344
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
MotivationConsumer Behavior
Algorithms and Methods for Distributed Storage Networks
13
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Consumer Usage
14
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Content Size
15
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Storage Hierarchy
‣ Primary storage• Processors registers• Processor cache• RAM
‣ Secondary storage• Hard disks• Solid state disks• CD, DVD
‣ Tertiary storage • tape libraries• optical jukeboxes
16
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Characteristics of Storage
‣ Volatile — non-volatile memory• non-volatile: dynamic or static
‣ Read & write — Read only — Slow write, fast read
‣ Random access – Sequential access‣ Addressability
• location addressable• file addressable• content addressable
‣ Capacity‣ Performance
• Latency• Throughput
17
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Non-volatile Storage Technologies
‣ Punch cards (Hollerith) 1886-1950s‣ Magnetic tape data storage 1951-today‣ Hard disk drive 1956-today‣ Floppy disks 1970s-1990s‣ EEPROM (Electrically Erasable Programmable
Read-Only Memory) 1980-today• Flash memory
‣ Optical disc drive (read/write) 1997-today
18
Distributed Storage Networksand Computer ForensicsWinter 2011/12
Computer Networks and TelematicsUniversity of Freiburg
Christian Schindelhauer
Network Storage Types
‣ Direct attached storage (DAS)• traditional storage
‣ Network attached storage (NAS)• storage attached to another computer accessible at file level
over LAN or WAN‣ Storage area network (SAN)
• specialized network providing other computers with storage capacity with access on block-addressing level
‣ File area network (FAN)• systematic approach to organize file-related storage systems• organization wide high-level storage network
19
University of FreiburgTechnical FacultyComputer Networks and TelematicsWinter Semester 2011/12
Distributed Storage Networksand Computer Forensics1. Organization & OverviewChristian SchindelhauerAmir Alsbih