Download Chapter 8: Device Chapter 8: Device Management

Chapter 8: Mass-Storage Systems
„ Overview of Mass Storage
Chapter 8: Device
Management
„ I/O Hardware
Structure
„ Application I/O Interface
„ Disk Structure
„ Kernel I/O Subsystem
„ Disk Attachment
„ Transforming I/O Requests to
„ Disk Scheduling
Hardware Operations
„ Disk Management
„ STREAMS
„ Swap-Space Management
„ Performance
„ RAID Structure
„ Disk Attachment
„ Stable-Storage
g Implementation
p
„ Tertiary Storage Devices
„ Operating System Support
„ Performance Issues
Operating System Concepts –
8th
Silberschatz, Galvin and Gagne ©2009
Edition,
Operating System Concepts – 8th Edition
Objectives
„
and the resulting effects on the uses of the devices
„ Explain the performance characteristics of mass-storage devices
Magnetic disks provide bulk of secondary storage of modern computers
z
Drives rotate at 60 to 200 times per second
z
Transfer rate is rate at which data flow between drive and computer
z
Positioning time (random-access time) is time to move disk arm to desired
cylinder
li d ((seek
k titime)) and
d titime ffor d
desired
i d sector
t tto rotate
t t under
d the
th disk
di k h
head
d
(rotational latency)
z
Head crash results from disk head making contact with the disk surface
„ Discuss operating-system services provided for mass storage, including
RAID and HSM
„ Explore the structure of an operating system’s I/O subsystem
„ Discuss the principles of I/O hardware and its complexity
„ Provide details off the performance
f
aspects off I/O
/O hardware and software
f
8.3
Silberschatz, Galvin and Gagne ©2009
Overview of Mass Storage Structure
„ Describe the physical structure of secondary and tertiary storage devices
Operating System Concepts – 8th Edition
8.2
Silberschatz, Galvin and Gagne ©2009
That’s bad
„
Disks can be removable
„
Drive attached to computer
p
via I/O bus
z
Busses vary, including EIDE, ATA, SATA, USB, Fibre Channel, SCSI
z
Host controller in computer uses bus to talk to disk controller built into drive or
storage array
Operating System Concepts – 8th Edition
8.4
Silberschatz, Galvin and Gagne ©2009
Moving-head Disk Mechanism
Overview of Mass Storage Structure (Cont)
„ Magnetic tape
Operating System Concepts – 8th Edition
8.5
Silberschatz, Galvin and Gagne ©2009
z
W early
Was
l secondary-storage
d
t
medium
di
z
Relatively permanent and holds large quantities of data
z
Access time slow
z
Random access ~1000 times slower than disk
z
Mainly used for backup, storage of infrequently-used data, transfer
medium between systems
z
Kept in spool and wound or rewound past read-write head
z
Once data under head
head, transfer rates comparable to disk
z
20-200GB typical storage
z
Common technologies
g
are 4mm, 8mm, 19mm, LTO-2 and SDLT
Operating System Concepts – 8th Edition
Disk Structure
8.6
Silberschatz, Galvin and Gagne ©2009
Disk Attachment
„ Disk drives are addressed as large 1-dimensional arrays of logical blocks,
where the logical block is the smallest unit of transfer
„ Host-attached storage accessed through I/O ports talking to I/O busses
„ SCSI itself
it lf is
i a bus,
b
up tto 16 devices
d i
on one cable,
bl SCSI iinitiator
iti t requests
t
operation and SCSI targets perform tasks
„ The 1-dimensional array of logical blocks is mapped into the sectors of the
di k sequentially
disk
ti ll
z
Sector 0 is the first sector of the first track on the outermost cylinder
z
Mapping proceeds in order through that track, then the rest of the tracks
in that cylinder, and then through the rest of the cylinders from
outermost to innermost
Operating System Concepts – 8th Edition
8.7
Silberschatz, Galvin and Gagne ©2009
z
Each target can have up to 8 logical units (disks attached to device
controller
„ FC is high-speed serial architecture
z
Can be switched fabric with 24-bit
24 bit address space – the basis of storage
area networks (SANs) in which many hosts attach to many storage units
z
Can be arbitrated loop (FC-AL) of 126 devices
Operating System Concepts – 8th Edition
8.8
Silberschatz, Galvin and Gagne ©2009
Network-Attached Storage
„ Network-attached storage (NAS) is storage made available over a network
rather than over a local connection (such as a bus)
Storage Area Network
„ Common in large storage environments (and becoming more common)
„ Multiple
M lti l hosts
h t attached
tt h d to
t multiple
lti l storage
t
arrays - flexible
fl ibl
„ NFS and CIFS are common protocols
„ Implemented via remote procedure calls (RPCs) between host and storage
„ New iSCSI protocol uses IP network to carry the SCSI protocol
8.9
Operating System Concepts – 8th Edition
Silberschatz, Galvin and Gagne ©2009
Disk Scheduling
Operating System Concepts – 8th Edition
8.10
Silberschatz, Galvin and Gagne ©2009
Disk Scheduling (Cont)
„ The operating system is responsible for using hardware efficiently — for the
disk drives
drives, this means having a fast access time and disk bandwidth
„ Several algorithms exist to schedule the servicing of disk I/O requests
„ We
W illustrate
ill t t th
them with
ith a requestt queue (0
(0-199)
199)
„ Access time has two major components
z
Seek time is the time for the disk are to move the heads to the cylinder
containing the desired sector
z
Rotational latency is the additional time waiting for the disk to rotate the
desired sector to the disk head
98, 183, 37, 122, 14, 124, 65, 67
Head pointer 53
„ Minimize seek time
„ Seek time ≈ seek distance
„ Disk bandwidth is the total number of bytes transferred, divided by the total
time between the first request for service and the completion of the last
transfer
Operating System Concepts – 8th Edition
8.11
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.12
Silberschatz, Galvin and Gagne ©2009
FCFS
SSTF
Illustration shows total head movement of 640 cylinders
„ Selects the request with the minimum seek time from the current head
position
„ SSTF scheduling is a form of SJF scheduling; may cause starvation of
some requests
„ Illustration shows total head movement of 236 cylinders
Operating System Concepts – 8th Edition
8.13
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
SSTF (Cont)
8.14
Silberschatz, Galvin and Gagne ©2009
SCAN
„ The disk arm starts at one end of the disk, and moves toward the other end,
servicing requests until it gets to the other end of the disk
disk, where the head
movement is reversed and servicing continues.
„ SCAN algorithm Sometimes called the elevator algorithm
„ Illustration shows total head movement of 208 cylinders
Operating System Concepts – 8th Edition
8.15
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.16
Silberschatz, Galvin and Gagne ©2009
SCAN (Cont.)
C-SCAN
„ Provides a more uniform wait time than SCAN
„ The
Th h
head
d moves ffrom one end
d off th
the di
disk
k tto th
the other,
th servicing
i i requests
t
as it goes
z
When it reaches the other end, however, it immediately returns to the
beginning of the disk, without servicing any requests on the return trip
„ Treats the cylinders as a circular list that wraps around from the last cylinder
to the first one
Operating System Concepts – 8th Edition
8.17
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
C-SCAN (Cont)
8.18
Silberschatz, Galvin and Gagne ©2009
C-LOOK
„ Version of C-SCAN
„ Arm
A only
l goes as ffar as the
th last
l t requestt in
i each
h di
direction,
ti
th
then reverses
direction immediately, without first going all the way to the end of the disk
Operating System Concepts – 8th Edition
8.19
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.20
Silberschatz, Galvin and Gagne ©2009
C-LOOK (Cont)
Selecting a Disk-Scheduling Algorithm
„ SSTF is common and has a natural appeal
„ SCAN and
dC
C-SCAN
SCAN perform
f
b
better
tt ffor systems
t
th
thatt place
l
ah
heavy lload
d on
the disk
„ Performance depends on the number and types of requests
„ Requests for disk service can be influenced by the file-allocation method
„ The disk-scheduling algorithm should be written as a separate module of
the operating system
system, allowing it to be replaced with a different algorithm if
necessary
„ Either SSTF or LOOK is a reasonable choice for the default algorithm
8.21
Operating System Concepts – 8th Edition
Silberschatz, Galvin and Gagne ©2009
Disk Management
Operating System Concepts – 8th Edition
8.22
Silberschatz, Galvin and Gagne ©2009
Booting from a Disk in Windows 2000
„ Low-level formatting, or physical formatting — Dividing a disk into sectors
that the disk controller can read and write
„ To use a disk to hold files, the operating system still needs to record its own
data structures on the disk
z
Partition the disk into one or more groups of cylinders
z
Logical formatting or “making a file system”
z
To increase efficiency most file systems group blocks into clusters
Disk I/O done in blocks
File I/O done in clusters
„ Boot block initializes system
z
The bootstrap is stored in ROM
z
Bootstrap loader program
„ Methods such as sector sparing used to handle bad blocks
Operating System Concepts – 8th Edition
8.23
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.24
Silberschatz, Galvin and Gagne ©2009
Swap-Space Management
„ Swap-space — Virtual memory uses disk space as an extension of main
RAID Structure
„ RAID – multiple disk drives provides reliability via redundancy
memory
„ Swap-space can be carved out of the normal file system, or, more
„ Increases the mean time to failure
commonly, it can be in a separate disk partition
„ Swap-space management
„ Frequently combined with NVRAM to improve write performance
z
4.3BSD allocates swap space when process starts; holds text segment
(the program) and data segment
z
Kernel uses swap maps to track swap-space use
z
Solaris 2 allocates swap space only when a page is forced out of
physical memory, not when the virtual memory page is first
f
created
8.25
Operating System Concepts – 8th Edition
Silberschatz, Galvin and Gagne ©2009
„ RAID is arranged into six different levels
Operating System Concepts – 8th Edition
8.26
Silberschatz, Galvin and Gagne ©2009
RAID 0 – Non redundant
RAID (Cont)
„ คือการเอา harddisk มากกวา 1 ตัวมาตอรวมกันในลักษณะ non-redundant ซึง่
„ 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
z
Mirroring or shadowing (RAID 1) keeps duplicate of each disk
z
Striped mirrors (RAID 1+0) or mirrored stripes (RAID 0+1) provides high
performance and high reliability
z
Block interleaved parity (RAID 4, 5, 6) uses much less redundancy
RAID 0 นี้มีจุดประสงค เพื่อทีจ่ ะเพิ่มความเร็วในการอาน/เขียนขอมูล harddisk โดยตรง ไมมี
การเก็บขอมูลสํารอง ดังนัน้ ถาฮารดดิสกตัวใดตัวหนึ่งเกิดเสียหาย ก็จะสงผลใหขอมูลทั้งหมดไม
สามารถใชงานไดทันที
„ จุดเดนของ RAID 0 คือความเร็วในการเขาถึงขอมูล แตขอเสียก็คือหาก harddisk ตัวใดตัว
หนึ่งเสียหาย จะสงผลกับขอมูลทั้งระบบทันที
„ RAID within a storage
g array
y can still fail if the array
y fails, so automatic
replication of the data between arrays is common
„ Frequently, a small number of hot-spare disks are left unallocated,
automatically replacing a failed disk and having data rebuilt onto them
Operating System Concepts – 8th Edition
8.27
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.28
Silberschatz, Galvin and Gagne ©2009
RAID 1: Mirrored Disks
„ RAID 1 (disk mirroring) ประกอบไปดวย harddisk 2 ตัวที่เก็บขอมูลเหมือนกันทุก
ประการ เสมือนการสํารองขอมูล หาก harddisk ตัวใดตัวหนึง่ เกิดเสียหาย ระบบก็ยังสามารถดึงขอมูล
จาก harddisk อีกตัวหนึ่งมาใชงานไดตามปกติ
„ สําหรับ RAID controller ที่ถูกออกแบบมาเปนอยางดีแลว การเขียนขอมูลลง harddisk 2 ตัว
ในเวลาเดียวกัน จะใชเวลาพอๆ กับการเขียนขอมูลลง harddisk ตัวเดียว ในขณะที่เวลาในการอานก็จะ
นอยลง เพราะ RAID controller จะเลือกอานขอมูลจาก harddisk ตัวไหนก็ได โดยหากมี
คําสัง่ ใหอานขอมลล 2 ชุชดในเวลาเดี
คาสงใหอานขอมู
ดในเวลาเดยวกน
ยวกัน ตว
ตัว RAID controller กสามารถประมวลผลคาสงเพอ
ก็สามารถประมวลผลคําสั่งเพื่อ
อานขอมูลจาก harddisk ตัวหนึง่ และคําสั่งอีกชุดนึงจาก harddisk อีกตัวนึงก็ได
RAID 2: Error-Correcting Codes
„ RAID 2 นี้ ขอมูลทั้งหมดจะถูกตัดแบงเพื่อจัดเก็บลง harddisk แตละตัวใน disk array
„ โดยมี harddisk ตัวหนึ่งเก็บขอมูลที่ใชตรวจสอบและแกไขขอผิดพลาด (ECC - Error Checking and
Correcting) ถาเกิดการปรากฏวา harddisk ตัวใดตัวหนึ่งเสียหาย ระบบก็จะสามารถสรางขอมูลทั้งหมดใน
harddisk ตัวนั้นขึ้นมาไดใหม โดยอาศัยขอมูลจาก harddisk ตัวอื่นๆ และจากคา ECC ที่เก็บเอาไว ซึง่ การทํา
ECC นี้สง ผลให harddisk ทั้งระบบตองทํางานคอนขางมากทีเดียว และ RAID 2 นั้นจะเห็นไดวาตองใช
harddisk จํานวนมากในการเก็บคา ECC ซึง่ ทําใหคอนขางสิ้นเปลือง
„ Hamming code is an algorithm that adds extra, redundant bits to the data
and
d iis th
therefore
f
able
bl tto correctt single-bit
i l bit errors and
dd
detect
t td
double-bit
bl bit
errors
„ จุจดเด
ดเดนของ
นของ RAID 1 คอความปลอดภยของขอมู
คือความปลอดภัยของขอมลล ไมเนนเรองประสทธภาพและความเรวเหมอนอยาง
ไมเนนเรื่องประสิทธิภาพและความเร็วเหมือนอยาง
RAID 0 แมวาประสิทธิภาพในการอานขอมูลของ RAID 1 จะสูงขึ้นก็ตาม
Operating System Concepts – 8th Edition
8.29
Silberschatz, Galvin and Gagne ©2009
RAID 3: Bit-interleaved parity
„ คลายกับ RAID 2 แตแทนทีจ่ ะตัดแบงขอมูลในระดับ bit เหมือน RAID 2 ก็จะตัดเก็บขอมูลใน
ระดับ byte
y แทนและการตรวจสอบและแกไขขอผิดพลาดของขอมููล จะใช p
parity
y แทนทีจ่ ะเปน
ECC ทําให RAID 3 มีความสามารถในการอานและเขียนขอมูลไดอยางรวดเร็ว เพราะมีการตอ
harddisk แตละตัวแบบ stripe และใช harddisk ที่เก็บ parity เพียงแคตัวเดียว
„ แตถานํา RAID 3 ไปใชในงานที่มีการสงผานขอมูลในจํานวนทีน
่ อ ยๆ ซึ่ง RAID 3 ตองกระจาย
Operating System Concepts – 8th Edition
8.30
Silberschatz, Galvin and Gagne ©2009
RAID 4: Block-Interleaved Parity
„ RAID 4 มีลักษณะโดยรวมเหมือนกับ RAID 3 ทุกประการ ยกเวนเรื่องการตัดแบงขอมูลทีท่ ํา
ในระดับ block แทนที่จะเปน bit หรือ byte ซึง่ ทําใหการอานขอมูลแบบ random ทําได
รวดเร็วกวา อยางไรก็ตาม ปญหาคอขวดที่กลาวใน RAID 3 ก็ยังมีโอกาสเกิดขึ้นเหมือนเดิม
„ ขอดีคือ ถามีดิสตใดเสีย สามารถกูขอมูลคืนโดยอาศัยตัว parity disk ได
ขอมูลไปทั่วทัง้ harddisk จะทําใหเกิดปญหาที่เรียกวา คอขวด ที่ harddisk ทีเ่ ก็บ parity
เพราะไม
ไ วาขอ มูลจะมีขี นาดใหญ
ใ ข นาดไหน
ไ RAID 3 ตอ งเสียี เวลาไปสร
ไป า งสว น parity
it ทังั้ สินิ้ ยิงิ่
ขอมูลมีขนาดเล็กๆ แต parity ตองสรางขึ้นตลอด ทําใหขอมูลถูกจัดเก็บเสร็จกอนการสราง parity
ทัง้ ระบบตองมารอใหสราง parity เสรจกอน
ทงระบบตองมารอใหสราง
เสร็จกอน จึจงจะทางานตอได
งจะทํางานตอได
„ RAID 3 เหมาะสําหรับใชในงานที่มีการสงขอมูลจํานวนมากๆ เชนงานตัดตอ Video
Operating System Concepts – 8th Edition
8.31
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.32
Silberschatz, Galvin and Gagne ©2009
RAID 5: Block-interleaved Distributed Parity
„ มีการตัดแบงขอมูลในระดับ block เชนเดียวกับ RAID 4 แตจะไมทําการแยก harddisk ตัว
ใดตัวหนึ่งเพื่อเก็บ parity
p y
„ ในการเก็บ parity ของ RAID 5 นัน
้ จะกระจาย parity ไปยัง harddisk ทุกตัว โดยปะปน
ไปกับขอมููลปกติ จึงชวยลดปญหาคอขวด ซึง่ เปนปญหาที่สําคัญใน RAID 3 และ RAID 4
„ คุณสมบัติอีกอันหนึง่ ทีน
่ าสนใจของ RAID 5 คือ เทคโนโลยี Hot Swap คือเราสามารถทําการ
RAID 6: P+Q Redundancy
„ RAID 6 อาศัยพื้นฐานการทํางานของ RAID 5 เกือบทุกประการ แตมีการเพิ่ม parity
block เขาไปอก
เขาไปอีก 1 ชุชดด เพอยอมใหเราทาการ
เพื่อยอมใหเราทําการ Hot Swap ไดพรอมกน
ไดพรอมกัน 2 ตว
ตัว (RAID 5 ทา
ทํา
การ Hot Swap ไดทีละ 1 ตัวเทานัน้ หาก harddisk มีปญหาพรอมกัน 2 ตัว จะทําใหเสีย
ทัง้ ระบบ)) เรียกวาเปนการเพิ่ม Fault Tolerance ใหกับระบบ
„ RAID 6 เหมาะกับงานที่ตองการความปลอดภัยและเสถียรภาพของขอมูลที่สงู มากๆ
เปลี่ยน harddisk ในกรณีที่เกิดปญหาไดในขณะที่ระบบยังทํางานอยูู เหมาะสําหรับงาน Server
ตางๆ ที่ตองทํางานตอเนื่อง
Operating System Concepts – 8th Edition
8.33
Silberschatz, Galvin and Gagne ©2009
RAID Levels
Operating System Concepts – 8th Edition
8.35
Operating System Concepts – 8th Edition
8.34
Silberschatz, Galvin and Gagne ©2009
RAID (0 + 1) and (1 + 0)
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.36
Silberschatz, Galvin and Gagne ©2009
Stable-Storage Implementation
Tertiary Storage Devices
„ Write-ahead log scheme requires stable storage
„ Low cost is the defining characteristic of tertiary storage
„ To implement stable storage:
„ Generally, tertiary storage is built using removable media
z
Replicate information on more than one nonvolatile storage media with
independent failure modes
z
Update information in a controlled manner to ensure that we can
recover the stable data after any failure during data transfer or recovery
8.37
Operating System Concepts – 8th Edition
Silberschatz, Galvin and Gagne ©2009
„ Common examples of removable media are floppy disks and CD-ROMs;
other types are available
Operating System Concepts – 8th Edition
Removable Disks
8.38
Silberschatz, Galvin and Gagne ©2009
Removable Disks (Cont.)
„ Floppy disk — thin flexible disk coated with magnetic material, enclosed
i a protective
in
t ti plastic
l ti case
z
Most floppies hold about 1 MB; similar technology is used for
removable disks that hold more than 1 GB
z
Removable magnetic disks can be nearly as fast as hard disks, but
theyy are at a g
greater risk of damage
g from exposure
p
„ A magneto-optic disk records data on a rigid platter coated with magnetic
material
z
Laser heat is used to amplify a large, weak magnetic field to record a bit
z
Laser light is also used to read data (Kerr effect)
z
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 – 8th Edition
8.39
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.40
Silberschatz, Galvin and Gagne ©2009
WORM Disks
Tapes
„ The data on read-write disks can be modified over and over
„ WORM (“Write
(“W it O
Once, R
Read
dM
Many Ti
Times”)
”) disks
di k can b
be written
itt only
l once
„ Thin aluminum film sandwiched between two glass or plastic platters
„ To write a bit,
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
„ 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
z
stacker – library that holds a few tapes
z
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
f active use
Operating System Concepts – 8th Edition
8.41
Silberschatz, Galvin and Gagne ©2009
Operating System Support
„ Major OS jobs are to manage physical devices and to present a virtual
machine abstraction to applications
Operating System Concepts – 8th Edition
8.42
Silberschatz, Galvin and Gagne ©2009
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
„ For hard disks, the OS provides two abstraction:
not not open a file on the tape, it opens the whole tape drive as a raw
d i
device
z
Raw device – an array of data blocks
z
File system – the OS queues and schedules the interleaved requests
from several applications
„ 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 it
Operating System Concepts – 8th Edition
8.43
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.44
Silberschatz, Galvin and Gagne ©2009
File Naming
Hierarchical Storage
g Management
g
(HSM)
(
)
„ The issue of naming files on removable media is especially difficult when we
want to write data on a removable cartridge on one computer
computer, and then use
the cartridge in another computer
„ Contemporary OSs generally leave the name space problem unsolved for
removable
bl media,
di and
dd
depend
d on applications
li ti
and
d users tto fifigure outt h
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 – 8th Edition
8.45
„ 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
z
Small and frequently used files remain on disk
z
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
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
Speed
8.46
Silberschatz, Galvin and Gagne ©2009
Speed (Cont)
„ Two aspects of speed in tertiary storage are bandwidth and latency
„ Bandwidth is measured in bytes per second
z
Sustained bandwidth – average data rate during a large transfer; # of
bytes/transfer time
Data rate when the data stream is actually flowing
z
Effective bandwidth – average over the entire I/O time, including
seek() or locate(), and cartridge switching
Drive’s overall data rate
„ Access latency – amount of time needed to locate data
z
Access time for a disk – move the arm to the selected cylinder
and wait for the rotational latency; < 35 milliseconds
z
Access on tape requires winding the tape reels until the selected
block reaches the tape head; tens or hundreds of seconds
z
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 – 8th Edition
8.47
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.48
Silberschatz, Galvin and Gagne ©2009
Reliability
Cost
„ A fixed disk drive is likely to be more reliable than a removable disk or tape
„ Main memory is much more expensive than disk storage
drive
„ The cost per megabyte of hard disk storage is competitive with magnetic
„ 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
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
y larger
g than the number of drives
8.49
Operating System Concepts – 8th Edition
Silberschatz, Galvin and Gagne ©2009
I/O Hardware
Operating System Concepts – 8th Edition
8.50
Silberschatz, Galvin and Gagne ©2009
A Typical PC Bus Structure
„ Incredible variety of I/O devices
„ Common
C
concepts
t
z
Port
z
Bus (daisy chain or shared direct access)
z
Controller (host adapter)
„ I/O instructions control devices
„ Devices have addresses, used by
z
Direct I/O instructions
z
Memory-mapped I/O
Operating System Concepts – 8th Edition
8.51
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.52
Silberschatz, Galvin and Gagne ©2009
Polling
Device I/O Port Locations on PCs (partial)
„ Determines state of device
z
command-ready
d
d
z
busy
z
Error
„ Busy-wait cycle to wait for I/O from device
8.53
Operating System Concepts – 8th Edition
Silberschatz, Galvin and Gagne ©2009
Interrupts
Operating System Concepts – 8th Edition
8.54
Silberschatz, Galvin and Gagne ©2009
Interrupt-Driven I/O Cycle
„ CPU Interrupt-request line triggered by I/O device
„ Interrupt handler receives interrupts
„ Maskable to ignore or delay some interrupts
„ Interrupt vector to dispatch interrupt to correct handler
z
Based on priority
z
Some nonmaskable
„ Interrupt mechanism also used for exceptions
Operating System Concepts – 8th Edition
8.55
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.56
Silberschatz, Galvin and Gagne ©2009
Direct Memory Access
Intel Pentium Processor Event-Vector Table
„ Used to avoid programmed I/O for large data movement
„ Requires DMA controller
„ Bypasses CPU to transfer data directly between I/O device and memory
Operating System Concepts – 8th Edition
8.57
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
Six Step Process to Perform DMA Transfer
8.58
Silberschatz, Galvin and Gagne ©2009
Application I/O Interface
„ I/O system calls encapsulate device behaviors in generic classes
„ Device-driver
D i di
llayer hid
hides diff
differences among I/O controllers
t ll
ffrom kkernell
„ Devices vary in many dimensions
Operating System Concepts – 8th Edition
8.59
Silberschatz, Galvin and Gagne ©2009
z
Character-stream
Character
stream or block
z
Sequential or random-access
z
Sharable or dedicated
z
Speed of operation
z
read-write, read only, or write only
Operating System Concepts – 8th Edition
8.60
Silberschatz, Galvin and Gagne ©2009
Characteristics of I/O Devices
A Kernel I/O Structure
Operating System Concepts – 8th Edition
8.61
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
Block and Character Devices
„ Block devices include disk drives
z
C
Commands
d iinclude
l d read,
d write,
it seek
k
z
Raw I/O or file-system access
z
Memory-mapped
Memory
mapped file access possible
8.62
Silberschatz, Galvin and Gagne ©2009
Network Devices
„ Varying enough from block and character to have own interface
„ Unix and Windows NT/9x/2000 include socket interface
z
Separates network protocol from network operation
z
Includes select() functionality
„ Character devices include keyboards, mice, serial ports
z
Commands include get(), put()
z
Libraries layered on top allow line editing
Operating System Concepts – 8th Edition
8.63
„ Approaches vary widely (pipes
(pipes, FIFOs
FIFOs, streams
streams, queues
queues, mailboxes)
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.64
Silberschatz, Galvin and Gagne ©2009
Clocks and Timers
Blocking and Nonblocking I/O
„ Provide current time, elapsed time, timer
„ Blocking - process suspended until I/O completed
„ Programmable interval timer used for timings, periodic interrupts
„ ioctl() (on UNIX) covers odd aspects of I/O such as clocks and timers
z
E
Easy
tto use and
d understand
d t d
z
Insufficient for some needs
„ Nonblocking - I/O call returns as much as available
z
User interface, data copy (buffered I/O)
z
Implemented via multi-threading
z
Returns quickly with count of bytes read or written
„ Asynchronous - process runs while I/O executes
Silberschatz, Galvin and Gagne ©2009
8.65
Operating System Concepts – 8th Edition
z
Difficult to use
z
I/O subsystem signals process when I/O completed
Operating System Concepts – 8th Edition
Two I/O Methods
8.66
Silberschatz, Galvin and Gagne ©2009
Kernel I/O Subsystem
„ Scheduling
z
S
Some
I/O requestt ordering
d i via
i per-device
d i queue
z
Some OSs try fairness
„ Buffering - store data in memory while transferring between devices
Synchronous
Operating System Concepts – 8th Edition
z
To cope with device speed mismatch
z
To cope with device transfer size mismatch
z
To maintain “copy semantics”
A
Asynchronous
h
8.67
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.68
Silberschatz, Galvin and Gagne ©2009
Device-status Table
Kernel I/O Subsystem
„ Caching - fast memory holding copy of data
z
Al
Always
jjustt a copy
z
Key to performance
„ Spooling - hold output for a device
z
If device can serve only one request at a time
z
i.e., Printing
„ Device reservation - p
provides exclusive access to a device
Operating System Concepts – 8th Edition
8.69
Silberschatz, Galvin and Gagne ©2009
z
System calls for allocation and deallocation
z
Watch out for deadlock
Operating System Concepts – 8th Edition
Error Handling
8.70
Silberschatz, Galvin and Gagne ©2009
I/O Protection
„ OS can recover from disk read, device unavailable, transient write failures
„ User process may accidentally or purposefully attempt to disrupt normal
operation via illegal I/O instructions
„ Most return an error number or code when I/O request fails
„ System error logs hold problem reports
Operating System Concepts – 8th Edition
8.71
z
All I/O instructions defined to be privileged
z
I/O must be performed via system calls
Silberschatz, Galvin and Gagne ©2009
Memory-mapped and I/O port memory locations must be protected
too
Operating System Concepts – 8th Edition
8.72
Silberschatz, Galvin and Gagne ©2009
Use of a System Call to Perform I/O
Kernel Data Structures
„ Kernel keeps state info for I/O components, including open file tables,
network connections
connections, character device state
„ Many, many complex data structures to track buffers, memory allocation,
“di t ” bl
“dirty”
blocks
k
„ Some use object-oriented methods and message passing to implement I/O
Operating System Concepts – 8th Edition
8.73
Silberschatz, Galvin and Gagne ©2009
UNIX I/O Kernel Structure
Operating System Concepts – 8th Edition
8.74
Silberschatz, Galvin and Gagne ©2009
I/O Requests to Hardware Operations
„ Consider reading a file from disk for a process:
Operating System Concepts – 8th Edition
8.75
Silberschatz, Galvin and Gagne ©2009
z
Determine device holding file
z
Translate name to device representation
z
Physically read data from disk into buffer
z
Make data available to requesting process
z
Return control to process
Operating System Concepts – 8th Edition
8.76
Silberschatz, Galvin and Gagne ©2009
Life Cycle of An I/O Request
STREAMS
„ STREAM – a full-duplex communication channel between a user-level
process and a device in Unix System V and beyond
„ A STREAM consists of:
- STREAM head interfaces with the user process
- driver end interfaces with the device
- zero or more STREAM modules between them.
„ Each module contains a read queue and a write queue
„ Message passing is used to communicate between queues
Operating System Concepts – 8th Edition
8.77
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
The STREAMS Structure
8.78
Silberschatz, Galvin and Gagne ©2009
Performance
„ I/O a major factor in system performance:
Operating System Concepts – 8th Edition
8.79
Silberschatz, Galvin and Gagne ©2009
z
Demands CPU to execute device driver, kernel I/O code
z
Context switches due to interrupts
z
Data copying
z
Network traffic especially stressful
Operating System Concepts – 8th Edition
8.80
Silberschatz, Galvin and Gagne ©2009
Intercomputer Communications
Improving Performance
„ Reduce number of context switches
„ Reduce
R d
d
data
t copying
i
„ Reduce interrupts by using large transfers, smart controllers, polling
„ Use DMA
„ Balance CPU, memory, bus, and I/O performance for highest throughput
Operating System Concepts – 8th Edition
8.81
Silberschatz, Galvin and Gagne ©2009
Operating System Concepts – 8th Edition
8.82
Silberschatz, Galvin and Gagne ©2009