Chapter 14: Disk Performance

Optimization Part 2

 Caching and Buffering

 Redundancy - RAID Structure
 Disk Attachment
 Stable-Storage Implementation
 Tertiary Storage Devices
 Operating System Issues
 Performance Issues
 Caching and Buffering

 Many systems maintain a disk cache buffer,

which is a region of main memory that the
operating system reserves for disk data. In
one context, the reserved memory acts as a
cache, allowing processes quick access to
data that would otherwise need to be fetched
from disk. The reserved memory also acts as
a buffer, allowing the operating system to
delay writing modified data until the disk
experiences a light load or until the disk head
is in a favorable position to improve I/O
performance.
 For example, an operating system may delay
writing modified data to disk to allow time for
multiple requests to contiguous locations to
enqueue, so that they can be serviced with
one I/O request.
Operating System Concepts 13.2 Silberschatz, Galvin and Gagne 2002
 Caching and Buffering

 The disk cache buffer presents several

challenges to operating system designers.
Because the size of the disk cache must be
limited to allow enough memory for active
processes, the designer must implement some
replacement strategy. The cache-replacement
question is similar to the page-replacement
question, and designers use many of the same
heuristics. Most commonly, designers choose
a strategy that replaces the least-recently
used item in the disk cache buffer.
 A second concern arises because disk
caching can lead to inconsistencies. Disk
cache buffers are maintained in volatile
memory, so if the system fails or loses power
while modified data is in the cache buffer,
those changes are lost.
Operating System Concepts 13.3 Silberschatz, Galvin and Gagne 2002
 Caching and Buffering
 To safeguard data against such problems, the
contents of the disk cache buffer are
periodically flushed to the hard disk; this
reduces the probability of data loss if the
system crashes.
 A system that employs write-back caching
does not write modified data to disk
immediately. Instead, the cache is written to
disk periodically, enabling the operating
system to batch multiple I/Os that are
serviced using a single request, which can
improve system performance.
 A system that employs write-through caching
writes data both to the disk cache buffer and
to disk each time cached data is modified.
This technique prevents the system from
batching requests, but reduces the possibility
of inconsistent data in the event of a system
crash
Operating System Concepts 13.4 Silberschatz, Galvin and Gagne 2002
 Defragmentation

 Defragmentation places file data in

contiguous blocks on disk, which improves
access times by reducing seek activity when
accessing sequential data.
 Disk reorganization places frequently or
heavily used data in favorable locations on
disk (e.g., midrange tracks for noncircular
scheduling strategies) to reduce average seek
times.
 There is a growing gap between processor
speed and disk speed. The reduced access
time due to data compression might outweigh
the overhead incurred by compressing and
decompressing data.

Operating System Concepts 13.5 Silberschatz, Galvin and Gagne 2002

 Defragmentation

 If programs generally exhibit spatial locality,

and the area of locality is not at the center of
the disk, then moving the head to the center
of the disk during each idle period requires a
wasteful seek to return to the disk's hot spot
when requests resume.
 A multiprogramming system can service
requests from multiple concurrent processes,
which may lead to several hot spots on the
disk. In this case, it is difficult to determine
which, if any, hot spot the read/write head
should move to when it is idle.

Operating System Concepts 13.6 Silberschatz, Galvin and Gagne 2002

 Redundancy

 There are many examples of redundancy being

employed in operating systems for a variety of
reasons.
 A common use of redundancy is creating
backups to ensure that if one copy of
information is lost, it can be restored.
 A multiprocessing system can have a pool of
identical processors available to assign to
processes and threads as needed.
 Such redundancy has several advantages.
Although the system could still function with
only a single processor, having the extra
processors yields better performance because
the processors can all work in parallel.

Operating System Concepts 13.7 Silberschatz, Galvin and Gagne 2002

 Redundancy
 It is also effective for fault tolerance—if one
processor fails, the system can continue
operating.
 RAID (Redundant Array of Independent
Disks)reduces access times to data on disks
by placing redundant copies of that data on
separate disks that may function in parallel.
 Redundant copies of the data can be placed
on different regions of the same disk, so that
the movement of the read/write head can be
minimized and the amount of rotational
movement of the disk before the data
becomes accessible, thus increasing
performance.
 Redundancy, of course, has its price. The
resources costs money and the hardware and
the software to support them can become
more complex. This 13.8
Operating System Concepts
is yet another example of
Silberschatz, Galvin and Gagne 2002
 RAID Structure

 RAID – multiple disk drives provides reliability

via redundancy.

 RAID is arranged into six different levels.

Operating System Concepts 13.9 Silberschatz, Galvin and Gagne 2002

 RAID (cont)

 Several improvements in disk-use techniques

involve the use of multiple disks working
cooperatively.

 Disk striping uses a group of disks as one

storage unit.

 RAID schemes improve performance and

improve the reliability of the storage system
by storing redundant data.
 Mirroring or shadowing keeps duplicate of each
disk.
 Block interleaved parity uses much less
redundancy.

Operating System Concepts 13.10 Silberschatz, Galvin and Gagne 2002

 RAID Levels

Operating System Concepts 13.11 Silberschatz, Galvin and Gagne 2002

 RAID (0 + 1) and (1 + 0)

Operating System Concepts 13.12 Silberschatz, Galvin and Gagne 2002

 Disk Attachment

 Disks may be attached one of two ways:

1. Host attached via an I/O port

2. Network attached via a network connection

Operating System Concepts 13.13 Silberschatz, Galvin and Gagne 2002

 Network-Attached Storage

Operating System Concepts 13.14 Silberschatz, Galvin and Gagne 2002

 Storage-Area Network

Operating System Concepts 13.15 Silberschatz, Galvin and Gagne 2002

 Stable-Storage Implementation

 Write-ahead log scheme requires stable

storage.

 To implement stable storage:

 Replicate information on more than one
nonvolatile storage media with independent
failure modes.
 Update information in a controlled manner to
ensure that we can recover the stable data after
any failure during data transfer or recovery.

Operating System Concepts 13.16 Silberschatz, Galvin and Gagne 2002

 Tertiary Storage Devices

 Low cost is the defining characteristic of

tertiary storage.

 Generally, tertiary storage is built using

removable media

 Common examples of removable media are

floppy disks and CD-ROMs; other types are
available.

Operating System Concepts 13.17 Silberschatz, Galvin and Gagne 2002

 Removable Disks

 Floppy disk — thin flexible disk coated with

magnetic material, enclosed in a protective
plastic case.

 Most floppies hold about 1 MB; similar

technology is used for removable disks that hold
more than 1 GB.
 Removable magnetic disks can be nearly as fast
as hard disks, but they are at a greater risk of
damage from exposure.

Operating System Concepts 13.18 Silberschatz, Galvin and Gagne 2002

 Removable Disks (Cont.)

 A magneto-optic disk records data on a rigid

platter coated with magnetic material.
 Laser heat is used to amplify a large, weak
magnetic field to record a bit.
 Laser light is also used to read data (Kerr
effect).
 The magneto-optic head flies much farther from
the disk surface than a magnetic disk head, and
the magnetic material is covered with a
protective layer of plastic or glass; resistant to
head crashes.

 Optical disks do not use magnetism; they

employ special materials that are altered by
laser light.

Operating System Concepts 13.19 Silberschatz, Galvin and Gagne 2002

 WORM Disks

 The data on read-write disks can be modified

over and over.
 WORM (“Write Once, Read Many Times”) disks
can be written only once.
 Thin aluminum film sandwiched between two
glass or plastic platters.
 To write a bit, the drive uses a laser light to
burn a small hole through the aluminum;
information can be destroyed by not altered.
 Very durable and reliable.
 Read Only disks, such ad CD-ROM and DVD,
com from the factory with the data pre-
recorded.

Operating System Concepts 13.20 Silberschatz, Galvin and Gagne 2002

 Tapes

 Compared to a disk, a tape is less expensive

and holds more data, but random access is
much slower.
 Tape is an economical medium for purposes
that do not require fast random access, e.g.,
backup copies of disk data, holding huge
volumes of data.
 Large tape installations typically use robotic
tape changers that move tapes between tape
drives and storage slots in a tape library.
 stacker – library that holds a few tapes
 silo – library that holds thousands of tapes
 A disk-resident file can be archived to tape for
low cost storage; the computer can stage it
back into disk storage for active use.

Operating System Concepts 13.21 Silberschatz, Galvin and Gagne 2002

 Operating System Issues

 Major OS jobs are to manage physical devices

and to present a virtual machine abstraction
to applications

 For hard disks, the OS provides two

abstraction:
 Raw device – an array of data blocks.
 File system – the OS queues and schedules the
interleaved requests from several applications.

Operating System Concepts 13.22 Silberschatz, Galvin and Gagne 2002

 Application Interface

 Most OSs handle removable disks almost

exactly like fixed disks — a new cartridge is
formatted and an empty file system is
generated on the disk.
 Tapes are presented as a raw storage medium,
i.e., and application does not not open a file
on the tape, it opens the whole tape drive as a
raw device.
 Usually the tape drive is reserved for the
exclusive use of that application.
 Since the OS does not provide file system
services, the application must decide how to
use the array of blocks.
 Since every application makes up its own
rules for how to organize a tape, a tape full of
data can generally only be used by the
program that created
Operating System Concepts
it.
13.23 Silberschatz, Galvin and Gagne 2002
 Tape Drives

 The basic operations for a tape drive differ

from those of a disk drive.
 locate positions the tape to a specific logical
block, not an entire track (corresponds to
seek).
 The read position operation returns the logical
block number where the tape head is.
 The space operation enables relative motion.
 Tape drives are “append-only” devices;
updating a block in the middle of the tape also
effectively erases everything beyond that
block.
 An EOT mark is placed after a block that is
written.

Operating System Concepts 13.24 Silberschatz, Galvin and Gagne 2002

 File Naming

 The issue of naming files on removable media

is especially difficult when we want to write
data on a removable cartridge on one
computer, and then use the cartridge in
another computer.
 Contemporary OSs generally leave the name
space problem unsolved for removable media,
and depend on applications and users to
figure out how to access and interpret the
data.
 Some kinds of removable media (e.g., CDs) are
so well standardized that all computers use
them the same way.

Operating System Concepts 13.25 Silberschatz, Galvin and Gagne 2002

 Hierarchical Storage Management
(HSM)

 A hierarchical storage system extends the

storage hierarchy beyond primary memory and
secondary storage to incorporate tertiary
storage — usually implemented as a jukebox
of tapes or removable disks.
 Usually incorporate tertiary storage by
extending the file system.
 Small and frequently used files remain on disk.
 Large, old, inactive files are archived to the
jukebox.
 HSM is usually found in supercomputing
centers and other large installations that have
enormous volumes of data.

Operating System Concepts 13.26 Silberschatz, Galvin and Gagne 2002

 Speed

 Two aspects of speed in tertiary storage are

bandwidth and latency.

 Bandwidth is measured in bytes per second.

 Sustained bandwidth – average data rate during
a large transfer; # of bytes/transfer time.
Data rate when the data stream is actually
flowing.
 Effective bandwidth – average over the entire I/O
time, including seek or locate, and cartridge
switching.
Drive’s overall data rate.

Operating System Concepts 13.27 Silberschatz, Galvin and Gagne 2002

 Speed (Cont.)
 Access latency – amount of time needed to
locate data.
 Access time for a disk – move the arm to the
selected cylinder and wait for the rotational
latency; < 35 milliseconds.
 Access on tape requires winding the tape
reels until the selected block reaches the
tape head; tens or hundreds of seconds.
 Generally say that random access within a
tape cartridge is about a thousand times
slower than random access on disk.
 The low cost of tertiary storage is a result
of having many cheap cartridges share a
few expensive drives.
 A removable library is best devoted to the
storage of infrequently used data, because
the library can only satisfy a relatively small
number of I/O requests per hour.
Operating System Concepts 13.28 Silberschatz, Galvin and Gagne 2002
 Reliability

 A fixed disk drive is likely to be more reliable

than a removable disk or tape drive.

 An optical cartridge is likely to be more

reliable than a magnetic disk or tape.

 A head crash in a fixed hard disk generally

destroys the data, whereas the failure of a
tape drive or optical disk drive often leaves
the data cartridge unharmed.

Operating System Concepts 13.29 Silberschatz, Galvin and Gagne 2002

 Cost

 Main memory is much more expensive than

disk storage

 The cost per megabyte of hard disk storage is

competitive with magnetic tape if only one
tape is used per drive.

 The cheapest tape drives and the cheapest

disk drives have had about the same storage
capacity over the years.

 Tertiary storage gives a cost savings only

when the number of cartridges is considerably
larger than the number of drives.

Operating System Concepts 13.30 Silberschatz, Galvin and Gagne 2002