Chapter 12: Mass-Storage Structure

1
Chapter 12 Mass-Storage Structure

• Disk attachment
• SSD (Solid state drive) vs. HDD (Hard disk drive)
• Disk I/O scheduling
• Disk management
• Formatting/partitioning

2
Resources

• Kernel Korner - I/O Schedulers
• http//www.linuxjournal.com/article/6931
• Describes soft deadline disk head scheduling and
  anticipatory scheduling in Linux 2.6 kernel
• See also the CFQ scheduler (Complete Fair Queuing
  Scheduler)
• http//en.wikipedia.org/wiki/CFQ

3
Disk Attachment

• Host-attached storage
• Storage accessed on host through local I/O port
• Uses hardware bus and host controller (e.g., IDE,
  ATA, SATA, FireWire, USB, SCSI, FC Fiber
  Channel)
• Network-attached storage
• Special purpose storage system attached remotely
  over a data network
• Clients access severs via RPC (e.g., NFS for Unix
  systems or CIFS for Windows systems)
• Storage-area network
• Private network (using storage protocols)
  connecting file servers and storage units

4
SSD (Solid State Drive)

• SSD provides random access to blocks of device
• Uniform access time to different blocks
• HDD does not provide random access
• Different access times to data blocks depending
  on where head is presently and location of goal
  block

5
Erase and SSD

• Erase Writing to the SSD must be preceded by a
  block erase (sets all bits to 1)
• Lifetime Measured in the number of erase cycles
• Wear leveling Firmware or operating system
  drivers must balance the numbers of erase cycles
  done on each block so that device does not fail
  prematurely

6
SSD Performance

• Single chip has relatively high latencies (SLC
  NAND)
• 25 µs to fetch (read) a 4K page from the array
  to the IO buffer
• 250 µs to commit (write) a 4K page from the IO
  buffer to the array
• 2 ms to erase a 256 kiB block
• With parallel chip operation, 250 MB/s effective
  read/write speeds

http//en.wikipedia.org/wiki/Solid-state_drive
7
hdparm command

• Measuring disk performance yourself on Linux
• http//www.cyberciti.biz/tips/how-fast-is-linux-sa
  ta-hard-disk.html
• Old disk system 57.9 MB/sec

8
Figure 12.1 Moving-head disk mechanism (HDD)
9
http//www.cprince.com/courses/cs5631spring09/news
/item1.pdf (Presentation by Yizhao Zhuang)
10
HDD (Hard Disk Drive) Terms

• Access (positioning) time
• Time to start to transfer data
• Components Seek time, rotational latency
• Seek time time for disk arm to move heads to the
  right cylinder
• Rotational latency time for disk to rotate to
  the desired sector
• Transfer rate
• Sustained bandwidth (p. 539) average data
  transfer rate during a large transfer that is
  the, number of bytes divided by transfer time
• data rate without positioning time
• Effective bandwidth (p. 507 539) average
  transfer rate including positioning time

11
HDD Scheduling

• Modern disks are addressed as a large array of
  blocks
• a disk address is a block number
• Block number is converted to a old style disk
  address (i.e., cylinder, head, sector) by the
  disk device firmware
• Generally, increasing block number implies
  physically adjacent sectors (see also
  http//www.linuxjournal.com/article/6931), and
  movement to inner cylinders of disk

12
Disk Requests

• Typically a disk device has a queue with several
  pending requests
• One issue How can we schedule requests to (a)
  improve system performance and throughput, and
  (b) reduce request latency?
• Some request ordering possibilities
• Prioritize disk requests by type
• E.g., page faults get priority over a processes
  file I/O
• Process priority could play a role
• Reduce access time by changing order of requests
• Just consider seek time

13
Some Algorithms To Reduce Seek Time
14
Scheduling Examples

• Assume
• List (queue) of track requests
• 98, 183, 37, 122, 14, 124, 65, 67
• Left most is first request
• R/W head starts at track 53
• Tracks range from 0-199
• Only a single platter
• Question
• Given a specific scheduling algorithm what is the
  total head movement required for a list of track
  requests?

15
FCFS

• Assume
• List (queue) of track requests
• 98, 183, 37, 122, 14, 124, 65, 67
• Left most is first request
• R/W head starts at track 53
• Tracks range from 0-199
• Only a single platter
• 1) Show the head movement over the disk surface
• 2) What is the total head movement?

16
FCFS
98-5345 183-9885 183-37146 122-3785 122-14108
124-14110 124-6559 67-652
Total head movement 640 tracks
17
FCFS Performance Issues

• It is possible for disk positions far from a
  current area of activity to be starved
  indefinitely?
• No. With FCFS, incoming (new) requests will
  eventually get processed.
• We are not reordering requests here.
• The algorithm is intrinsically fair, but it
  generally does not generally provide the fastest
  service.

18
SSTF Shortest Seek Time First

• Assume
• List (queue) of track requests
• 98, 183, 37, 122, 14, 124, 65, 67
• Left most is first request
• R/W head starts at track 53
• Tracks range from 0-199
• Only a single platter
• 1) Show the head movement over the disk surface
• 2) What is the total head movement?

19
SSTF example
Total head movement 236 tracks
20
SSTF
Total head movement 236 tracks
21
SSTF Performance Issues

• It is possible for disk positions far from a
  current area of activity to be starved
  indefinitely?
• YES If new requests keep arriving for the
  current area
• Prefer a method that we know will not starve
  requests
• Also, while better than FCFS (236 for SSTF vs.
  640 for FCFS), less head movement can be obtained
• E.g., service order 53, 37, 14, 65, 67, 98, 122,
  124, 183 has total head movement of 208 tracks

22
SCAN

• Algorithm Move head from end to end (has a
  current direction), servicing requests as you
  come to them
• Assume
• List (queue) of track requests
• 98, 183, 37, 122, 14, 124, 65, 67
• Left most is first request
• R/W head starts at track 53
• Tracks range from 0-199
• Only a single platter
• Start direction down (to lower numbered tracks)
• 1) Show the head movement over the disk surface
• 2) What is the total head movement?
• From start position to servicing last request in
  list

23
SCAN
Total head movement 53 183 236
24
SCAN Performance Issues

• It is possible for disk positions far from a
  current area of activity to be starved
  indefinitely?

25
C-SCAN

• Algorithm Only service requests in one direction
  (circular SCAN)
• Assume
• List (queue) of track requests
• 98, 183, 37, 122, 14, 124, 65, 67
• Left most is first request
• R/W head starts at track 53
• Tracks range from 0-199
• Only a single platter
• Start direction up (to higher numbered tracks)
• Service requests (only) in up direction
• 1) Show the head movement over the disk surface
• 2) What is the total head movement?

26
C-SCAN
Total head movement 199-5337183 ( wrap)
27
C-LOOK

• Algorithm Only service requests in one direction
  and turn at last request in direction (circular
  LOOK)
• Assume
• List (queue) of track requests
• 98, 183, 37, 122, 14, 124, 65, 67
• Left most is first request
• R/W head starts at track 53
• Tracks range from 0-199
• Only a single platter
• Start direction up (to higher numbered tracks)
• Service requests in up direction
• 1) Show the head movement over the disk surface
• 2) What is the total head movement?

28
C-LOOK
Total head movement (183-53)(37-14)153 ( wrap)
29
Algorithm Selection Issues

• Starvation
• FCFS, SCAN and LOOK algorithms will not starve
  requests
• SSTF can starve requests
• Type of request activity
• Some requests will be to similar tracks of disk
• E.g., Requests by a single process for a file
  allocated using contiguous file allocation
• FCFS, SSTF Should do well with these clustered
  requests
• Some requests will be on widely different tracks
  of disk
• E.g., Indexed, linked file allocation requests
  by different processes
• May be best to have algorithm that distributes
  requests uniformly across surface of disk (e.g.,
  SCAN, LOOK)

30
Ch 12.5 Disk Management

• Disk formatting
• Swap space
• Boot blocks Booting

31
Ch 12.5.1 Disk Formatting

• Three steps
• Low-level formatting, partitioning, logical
  formatting
• Low-level formatting
• Fills disk sectors with special data structure
  that is used by disk I/O controller
• Header, data area (usually 512 bytes), trailer
• Header trailer information used by disk
  controller hardware-- e.g., sector number and ECC
  (error correcting code)
• Creates map of bad blocks
• Reserves spare sectors for bad block repair
• Partitioning
• Grouping disk into one or more groups of
  cylinders to be treated as separate logical disks
• Each partition can have its own file system type

32
Logical Formatting (Per Partition)

• Creation of organization for different types of
  file systems, swap space, database application
  etc.
• E.g., directory structures, free space lists
• Boot partitions
• Boot block code to load kernel of O/S
• It is also possible to leave the partition
  without logical formatting
• With no file system disk data structures
• Partition accessed as an array of blocks
• raw disk, and raw I/O
• Raw I/O bypasses file system services such as
  buffer cache, file locking, prefetching, space
  allocation, file names, and directories

33
Swap Space

• Virtual memory
• Will use regular file system for at least reading
  code and data of program
• Swap space may be on separate device or multiple
  devices
• Partition formatted for swap space can give
  higher virtual memory performance Why?
• With a separate partition, swap space manager
  allocates/deallocates blocks
• Doesnt use functions of normal file system
• Maintain map(s) of swap space usage
• May handle text (code) data differently
• E.g., Solaris 1 dont write text to swap disk
  if page replacement needs page, next time just
  re-read from normal file system dirty data pages
  still written to swap disk as needed

34
Boot Process

• Typically a three phase process
• ROM bootstrap
• Boot block loader
• Running operating system kernel
• ROM bootstrap program
• Initial program code that starts loading
  operating system
• Typically a relatively simple loader program
  stored in ROM reads the disk boot block(s) into
  RAM, and starts running this boot block code
• Boot block code
• Actually does loading of operating system kernel
  from boot partition

35
Handling Bad Blocks

• Can handle at free blocks level in O/S
• Free list of blocks can be used to indicate these
  blocks cannot be used
• Device controller
• Can maintain list of bad blocks
• Initialized during low-level formatting at
  factory
• Low-level formatting also sets aside spare
  sectors, not visible to O/S
• Sector sparing or forwarding can be used to
  manage new bad blocks
• Sparing or forwarding Controller uses a spare
  sector when it gets a request for a sector that
  has a bad block
• When block is remapped, controller uses a spare
  sector from the same cylinder if possible
• Sector slipping Moving a sequence of blocks
  re-mapping them (p. 519)

36
RAID Redundant Array of Independent/Inexpensive
Disks

• Multiple disk techniques for
• Improving throughput and/or response time
• Multiple disks, parallelize disk operations
• Improving data reliability
• Decrease mean time to failure (MTBF)
• If one disk fails, can replace with another and
  go on quickly, with easy recovery
• Ideas
• Data striping Methods of improving throughput
• Putting part of data on disk A, part on disk B,
  part on
• Bit striping e.g., 8 disks, one bit of each byte
  on each disk
• Block striping n disks, block i goes on disk (i
  mod n) 1
• Methods of improving reliability
• Mirroring (a.k.a. shadowing) duplicating all
  data
• Parity, ECC (error correcting codes)

37
END!
38
(No Transcript)