DSMDR

8/13/2019 DSMDR

1/40

Intelligent Disk Subsystem 1

8/13/2019 DSMDR

2/40

After completing this chapter, you will be able to: Internal Disk subsystem structure

Hard disk and internal I/O channels

JBOD: Just a Bunch Of Disk Storage virtualization using RAID

Different RAID levels

Caching: Acceralaration of Hard Disk Access

Intelligence and Availability of Disk Sub system

Intelligent Disk Subsystem - 2

8/13/2019 DSMDR

3/40

Small storage device is replaced by largestorage subsystem.

Performance limitation of small storagedevice.

An individual drive has a certain lifeexpectancy Measured in MTBF

Example If the MTBF of a drive is 750,000 hours, and there are

1000 drives in the array, then the MTBF of the arraybecomes 750,000 /1000, or 750 hours


8/13/2019 DSMDR

4/40

Storage consolidation.

High Availability.

High performance Instant copies

Remote mirroring at reasonable price


8/13/2019 DSMDR

5/40

Disk subsystem visualized as hard disk server. Servers are connected to a disk subsystem using

standard I/O techniques. SCSI: Small computer system interface

iSCSI: Internet SCSI

Fibre Channel (SAN)

Storage Pooling: Free disk space can be flexiblyassigned to each server connected.


8/13/2019 DSMDR

6/40


8/13/2019 DSMDR

7/40Intelligent Disk Subsystem - 7

The controller

increase the data availability

data access performance with

RAID.

facilitates copying services

instant copy and remote

mirroring.

uses a cache in an attempt toaccelerate read and write accesses

to the server.

8/13/2019 DSMDR


All servers share

the storage capacity

of a disk subsystem.

Each server can

be assigned freestorage more flexibly

as required.

8/13/2019 DSMDR


8/13/2019 DSMDR

10/40

Size of the hard disk used limits the maximumcapacity of the overall disk subsystem.

Maximum Performance Vs Maximum capacity

Maximum disk (C,I)

More disk more load in terms of R/W operationsand I/O channels

Factors impacting disk sub system or individualdisk: Size Speed Cache I/O communication medium.

Intelligent Disk Subsystem-

10

8/13/2019 DSMDR

11/40

For increased fault tolerance: I/O Channels techniques Active

Active/Passive

Active/Active (No Load sharing)

Active/ Active (Load Sharing).


11

8/13/2019 DSMDR

12/40Intelligent Disk Subsystem

-12

8/13/2019 DSMDR


-13

8/13/2019 DSMDR

14/40


14

8/13/2019 DSMDR

15/40

Disk Subsystem complexity No controller

RAID controller

Intelligent controller

JBOD: No internal controller.

Only full enclosure of disks

Normally 8 to 16 hard disk space. Every hard disk has a separate address space

Does not support RAID or any other virtualization


15

8/13/2019 DSMDR

16/40

Individual hard disk are slow and has less lifeexpectancy.

Stripping: Increases performance by strippingand

Mirroring: Improves fault tolerance byredundancy.

Parity: Provides data integrity RAID provides:

Increase capacity Higher availability Increased performance


16

8/13/2019 DSMDR

17/40


17

RAIDController

Hard Disks

LogicalArray

PhysicalArray

RAID Array

Host

8/13/2019 DSMDR

18/40


18

RAIDController

8/13/2019 DSMDR

19/40

Hardware (usually a specialized diskcontroller card) Controls all drives attached to it

Array(s) appear to host operating system as a

regular disk drive Provided with administrative software

Software Runs as part of the operating system

Performance is dependent on CPU workload Does not support all RAID levels


19

8/13/2019 DSMDR

20/40

0 Striped array with no fault tolerance 1 Disk mirroring with fault tolerance Nested RAID (i.e., 1 + 0, 0 + 1) 3 Parallel access array with dedicated parity

disk 4 Striped array with independent disks and a

dedicated parity disk 5 Striped array with independent disks and

distributed parity 6 Striped array with independent disks and

dual distributed parity


20

8/13/2019 DSMDR

21/40


21

Stripe 1

Stripe 2

Strips

Strip 1=64KB

Stripe=192KB

Strip

Stripe

Strip 2=64KB Strip 3=64KB

8/13/2019 DSMDR

22/40


22

1

9

5

2

10

6

3

117

0

Host

RAIDController

8/13/2019 DSMDR


Block 1 Block 1Block 1Block 0Block 0

Host

Block 0RAID

Controller

8/13/2019 DSMDR

24/40


24

Host

Block 5Block 4Block 2Block 1

Block 5

Block 4

Block 2

RAID 0

Block 1

RAID 1

Block 0Block 3RAID

Controller

8/13/2019 DSMDR

25/40


25

RAIDController

Block 5

Block 4

Block 2

RAID 0

Block 1

RAID 1

Block 5

Block 4

Block 2

Block 1

Block 5

Block 4

Block 2

Block 1

Host

8/13/2019 DSMDR

26/40


26

Host

Block 5

Block 5

Block 2

RAID 1Block 0Block 0

Block 2

RAID 0

Block 3Block 3RAID

Controller

Block 1

Block 1

Block 4

Block 4

8/13/2019 DSMDR

27/40


27

Host

RAIDController

RAID 1

Block 1

Block 1

RAID 0

Block 4

Block 4 Block 5

Block 5

Block 2

Block 2Block 1

Block 4

8/13/2019 DSMDR

28/40


28Parity Disk

1

9

5

3

117

0

0 1 2 3

4 5 6 7

4

6

1

7

18

Host

RAIDController

Parity calculation 4 + 6 + 1 + 7 = 18The middle drive fails:

4 + 6 + ? + 7 = 18

?= 18 4 6 7

? = 1

?

8/13/2019 DSMDR

29/40


29

Host

RAIDController

Block 1

Block 2

Block 3

P 0 1 2 3

Block 0Block

ParityGenerated

8/13/2019 DSMDR

30/40


30

Host

Block 0

P 0 1 2 3

P 4 5 6 7

RAIDController

P 0 1 2 3

Block 0Block 0

Block 1

Block 5

Block 2

Block 6

Block 3

ParityGenerated

Block 0

P 0 1 2 3

Block 4

Block 7

8/13/2019 DSMDR

31/40


31

Host

Block 0

P 0 1 2 3

Block 7

RAIDController

P 0 1 2 3

Block 0Block 4Block 0

Block 1

Block 5

Block 2

Block 6

Block 3

ParityGenerated

Block 0

P 0 1 2 3

Block 4

P 4 5 6 7P 4 5 6 7

Block 4

P 4 5 6 7

Block 4Parity

Generated

8/13/2019 DSMDR

32/40

Two disk failures in a RAID set leads to dataunavailability and data loss in single-parityschemes, such as RAID-3, 4, and 5

Increasing number of drives in an array andincreasing drive capacity leads to a higher

probability of two disks failing in a RAID set RAID-6 protects against two disk failures by

maintaining two parities Horizontal parity which is the same as RAID-5 parity Diagonal parity is calculated by taking diagonal sets of

data blocks from the RAID set members Even-Odd, and Reed-Solomon are two commonly

used algorithms for calculating parity in RAID-6


32

8/13/2019 DSMDR

33/40

RAIDMin

Disks

Storage

Efficiency %

Cost Read Performance Write Performance

0 2 100 Low

Very good for bothrandom and sequential

readVery good

1 2 50 HighGood

Better than a single disk

GoodSlower than a single

disk, as every write mustbe committed to two

disks

3 3

(n-1)*100/nwhere n=number of

disksModerate

Good for random readsand very good forsequential reads

Poor to fair for smallrandom writesGood for large,

sequential writes

45 3


disks

Moderate

Very good for randomreads

Good for sequential

reads

Fair for random writeSlower due to parity

overhead

Fair to good forsequential writes

6 4


disks

Moderatebut more

than RAID 5

Very good for randomreads

Good for sequentialreads

Good for small, randomwrites

(has write penalty)

1+0and

0+1

4 50 High Very good Good

Intelligent Disk Subsystem

RAID Comparison

8/13/2019 DSMDR

34/40


RAID Comparison

8/13/2019 DSMDR

35/40

Small (less than element size) write on RAID 5 Ep = E1 + E2 + E3 + E4 (XOR operations)

If parity is valid, then: Ep new = Ep old E4 old + E4 new (XOR operations) 2 disk reads and 2 disk writes

Parity Vs Mirroring Reading, calculating and writing parity segment introduces penalty to every write operation Parity RAID penalty manifests due to slower cache flushes Increased load in writes can cause contention and can cause slower read response times


35

Ep new

RAID Controller

2 XOR

+-= E4 oldEp old E4 new

P0 D1 D2 D3 D4

8/13/2019 DSMDR

36/40

Total IOPS at peak workload is 1200 Read/Write ratio 2:1

Calculate IOPS requirement at peak activityfor RAID 1/0

RAID 5


36

8/13/2019 DSMDR

37/40

Key points covered in this chapter: What RAID is and the needs it addresses

The concepts upon which RAID is built

Some commonly implemented RAID levels


37

8/13/2019 DSMDR

38/40

Click the attached file


8/13/2019 DSMDR

39/40

Click the attached file


8/13/2019 DSMDR

40/40

What is a RAID array? What benefits do RAID arrays provide?

What methods can be used to provide higherdata availability in a RAID array?

What is the primary difference between RAID3 and RAID 5?

What is advantage of using RAID 6?

What is a hot spare?

Documents

DSMDR