4 - Devices

Question 1

What are the two device types?

Answer 1

Shareable

• based on space sharing
• multiple requests

Non-shareable
- allocated to one program/request

Question 2

What are the components of device driver software

Answer 2

Init
Request Queuer
Queue
Request Dequeuer
---> Device Controller/Device
Transfer initiator
Interrupt Service ROutiner

Question 3

Define disk scheduling

Answer 3

Takes requests from the queue so as to optimise transfer of data

Question 4

How fast do magnetic drives rotate

Answer 4

60 to 200 times a second

Question 5

Transfer rate

Answer 5

speed at which data flows between drive and computer

Question 6

Access Time

Answer 6

Random access time is the time to move the disk arm to desired track (seek) and for the sector to rotate under the disk head (rotational latency)

Question 7

Head Crash

Answer 7

When the disk head makes contact with the disk surface

Question 8

FCFS Disk Scheduling

Answer 8

First COme First Served.

Order in which requests come

fair but not fast

Question 9

SSTF Disk Scheduling

Answer 9

Shortest Seek time first

Select the next closest request from the current pos.

May starve some requests that endlessly wait

Question 10

SCAN disk scheduling

Answer 10

Disk arm starts at end, moves towards other and services requests then reverses etc

elevator algorithm

Question 11

How does UNIX treat devices that can be written to or read from?

Answer 11

As special files (according to the inode)

Question 12

What is RAID?

Answer 12

Redundant Array of Independent Disks

Set of drives viewed as single logical unit

Question 13

What is RAID 0?

Answer 13

Striping

Distribute data over multiple disks in two forms:

• bit level
• block level

Question 14

How does bit level striping work?

Answer 14

Split the bits across multiple disks

• take 8 disks, bit i to disk i for each byte then access rate is eight times as large

Question 15

How does block level striping work?

Answer 15

Given n disks, block i of a file goes to disk (i mod n ) + 1

• block nums from 0 and disks from 1

Question 16

What are the pros and cons of RAID 0?

Answer 16

Pros
- Faster

Cons
- Any drive fails = lost data

Question 17

What is RAID 1?

Answer 17

Mirroring

improve reliability through redundancy

Copy data to two separate disks
- good error rec

• expensive
• writing slow
• reading improved

Question 18

What is the MTTF for mirroring?

Answer 18

If the mean time to failure of a single disk is 100,000 hours and the mean time to repair is 10 hours, the mean time to data loss is (100,000)^2 / (210) = 500106
hours, or 57,000 years!

Question 19

What are the various RAID levels?

Answer 19

• RAID 0 striping
• 1 mirroring
• 2 to 5 use parity bits to reduce disks required while still being redundant and higher performance through striping
• 6 uses parity and a different error code. Tolerates two disk failures