A Multi-Platter Disk

1
 Addressing and Geometry
• Externally, hard drives expose a large number of sectors
(blocks)
– Typically 512 or 4096 bytes
– Individual sector writes are atomic
– Multiple sectors writes may be interrupted (torn write)
• Drive geometry
– Sectors arranged into tracks
– A cylinder is a particular track on multiple platters
– Tracks arranged in concentric circles on platters
– A disk may have multiple, double-sided platters
• Drive motor spins the platters at a constant rate
– Measured in revolutions per minute (RPM) 2
 Geometry Example
Sector Three tracks

12
11 13
23 24 25
10 14
32
9 22 31 33 26 15
Rotation

8 21 30 34 27 0 Read head

7 20 29 35 16 1
28 Seeks across the
6 19
18
17 2 various tracks
5 3
4

Outer tracks hold

One platter
more data 3
 Redundant Array of Inexpensive Disks
• RAID: use multiple disks to create the illusion of a
large, faster, more reliable disk
• Externally, RAID looks like a single disk
– i.e. RAID is transparent
– Data blocks are read/written as usual
– No need for software to explicitly manage multiple disks or
perform error checking/recovery
• Internally, RAID is a complex computer system
– Disks managed by a dedicated CPU + software
– RAM and non-volatile memory
– Many different configuration options (RAID levels)

4
 RAID 0: Striping
• Key idea: present an array of disks as a single large disk
• Maximize parallelism by striping data cross all N disks

Disk 0 Disk 1 Disk 2 Disk 3

Data block = Random accesses
Sequential 0 1 2 3
512 bytes
accesses are naturally
4 5 6 7 spread over all
spread across
8 9 10 11 Stripe
drives
all drives
12 13 14 15

5
 Addressing Blocks
• How do you access specific data blocks?
– Disk = logical_block_number % number_of_disks
– Offset = logical_block_number / number_of_disks
• Example: read block 11
– 11 % 4 = Disk 3
– 11 / 4 = Physical Block 2 (starting from 0)
Disk 0 Disk 1 Disk 2 Disk 3

0 1 2 3
4 5 6 7
8 9 10 11
12 13 14 15 6
 Analysis of RAID 0
• Capacity: N
– All space on all drives can be filled with data
• Reliability: 0
– If any drive fails, data is permanently lost
• Sequential read and write: N * S
– Full parallelization across drives
• Random read and write: N * R
– Full parallelization across all drives

7
 RAID 1: Mirroring
• RAID 0 offers high performance, but zero error
recovery
• Key idea: make two copies of all data

Disk 0 Disk 1

0 0
1 1
2 2
3 3

8
 Disk 0 Disk 1 Disk 2 Disk 3 Analysis of RAID 1 (1)
0 1 0 1
2 3 2 3 Disk 0 Disk 1 Disk 2 Disk 3
4 5 4 5
6 7 6 7 skipped 0 1 0 1
Each 2 3 2 3
block is 4 5 4 5
wasted 6 7 6 7

• Capacity: N / 2 • Sequential write: (N / 2) * S

– Two copies of all data, thus half – Two copies of all data, thus
capacity half throughput
• Reliability: 1 drive can fail, • Sequential read: (N / 2) * S
sometime more – Half of the read blocks are
wasted, thus halving
– If you are lucky, N / 2 drives can fail throughput
without data loss
9
 Analysis of RAID 1 (2)
• Random read: N * R
– Best case scenario for RAID 1
– Reads can parallelize across all disks
• Random write: (N / 2) * R
– Two copies of all data, thus half throughput

Disk 0 Disk 1 Disk 2 Disk 3

0 1 0 1
2 3 2 3
4 5 4 5
6 7 6 7
10
 RAID 0+1 and 1+0 Examples
0+1 1+0
Mirror Stripe

Stripe Stripe Mirror Mirror

Disk 0 Disk 1 Disk 2 Disk 3 Disk 0 Disk 1 Disk 2 Disk 3

0 1 0 1 0 0 1 1
2 3 2 3 2 2 3 3
4 5 4 5 4 4 5 5
6 7 6 7 6 6 7 7

• Combines striping and mirroring

• Superseded by RAID 4, 5, and 6 11
 RAID 4: Parity Drive
Disk 0 Disk 1 Disk 2 Disk 3 Disk 4

0 1 2 3 P0 Disk N only stores

4 5 6 7 P1 parity information
8 9 10 11 P2 for the other N-1
12 13 14 15 P3 disks

Disk 0 Disk 1 Disk 2 Disk 3 Disk 4

0 0 1 1 0^0^1^1=0 Parity
0 1 0 0 0^1^0^0=1 calculated
1 1 1 1 1^1^1^1=0 using XOR
0 1 1 1 0^1^1^1=1

12
 Analysis of RAID 4
• Capacity: N – 1
– Space on the parity drive is lost
• Reliability: 1 drive can fail
• Sequential Read and write: (N – 1) * S
– Parallelization across all non-parity blocks
• Random Read: (N – 1) * R
– Reads parallelize over all but the parity drive
• Random Write: R / 2
– Writes serialize due to the parity drive
– Each write requires 1 read and 1 write of the parity drive, thus
R/2
13
 RAID 5: Rotating Parity

Disk 0 Disk 1 Disk 2 Disk 3 Disk 4

0 1 2 3 P0 Parity blocks are

5 6 7 P1 4 spread over all N
10 11 P2 8 9 disks
15 P3 12 13 14

Disk 0 Disk 1 Disk 2 Disk 3 Disk 4

0 0 1 1 0^0^1^1=0
1 0 0 0^1^0^0=1 0
1 1 1^1^1^1=0 1 1
1 0^1^1^1=1 0 1 1
14
 Analysis of Raid 5
• Capacity: N – 1 [same as RAID 4]
• Reliability: 1 drive can fail [same as RAID 4]
• Sequential Read and write: (N – 1) * S [same]
– Parallelization across all non-parity blocks
• Random Read: N * R [vs. (N – 1) * R]
– Unlike RAID 4, reads parallelize over all drives
• Random Write: N / 4 * R [vs. R / 2 for RAID 4]
– Unlike RAID 4, writes parallelize over all drives
– Each write requires 2 reads and 2 write, hence N / 4

15
 Comparison of RAID Levels

• N – number of drives • R – random access speed

• S – sequential access speed • D – latency to access a single disk
RAID 0 RAID 1 RAID 4 RAID 5
Capacity N N/2 N–1 N–1
Reliability 0 1 (maybe N / 2) 1 1
Sequential Read N*S (N / 2) * S (N – 1) * S (N – 1) * S
Throughput

Sequential Write N*S (N / 2) * S (N – 1) * S (N – 1) * S

Random Read N*R N*R (N – 1) * N*R
R
Random Write N*R (N / 2) * R R/2 (N / 4) * R
Latency

Read D D D D
Write D D 2*D 2*D

16
 RAID 6
Disk 0 Disk 1 Disk 2 Disk 3 Disk 4

0 1 2 P00 P01 Two parity blocks

4 5 P10 P11 3 per stripe
8 P20 P21 6 7
P30 P31 9 10 11

• Any two drives can fail

• N – 2 usable capacity
• No overhead on read, significant overhead on write
• Typically implemented using Reed-Solomon codes
17
 Choosing a RAID Level
• Best performance and most capacity?
– RAID 0
• Greatest error recovery?
– RAID 1 (1+0 or 0+1) or RAID 6
• Balance between space, performance, and
recoverability?
– RAID 5

18