file structuresnu-oopsla lab1 chap3. secondary storage and system software 서울대학교...

File Structure SNU-OOPSLA Lab 1

Chap3.Chap3. Secondary Storage Secondary Storage and System Software and System Software

서울대학교 컴퓨터공학과객체지향시스템연구실(SNU-OOPSLA-LAB)

김 형 주 교수

File Structures by Folk, Zoellick, and Ricarrdi


Chapter ObjectivesChapter Objectives Describe the organization of typical disk drives, including basic units of

organization and their relationships Identity and describes the factors affecting disk access time, and describe

methods for estimating access times and space requirements Describe magnetic tapes, identify some tape applications, and investigate

the implications of block size on space requirements and transmission speeds

Identify fundamental differences between media and criteria that can be used to match the right medium to an application

Describe in general terms the events that occur when data is transmitted between a program and a secondary storage device

Introduce concepts and techniques of buffer management Illustrate many of the concepts introduced in the chapter, especially system

software concepts, in the context of UNIX


ContentsContents3.1 Disks

3.2 Magnetic Tape

3.3 Disk versus Tape

3.4 Introduction to CD-ROM

3.5 Physical Organization of CD-ROM

3.6 CD-ROM Strengths and Weaknesses

3.7 Storage as a Hierarchy

3.8 A Journey of a Byte

3.9 Buffer Management

3.10 I/O in UNIX


DisksDisks Random Access Device Disk device - DASD(Direct Access Storage Devices)

Magnetic disk hard disk - most common floppy disk - inexpensive, little capacity magnetic tape - serial access Iomega Zip (100 M byte), Jaz (1 G byte)

Nonmagnetic disk optical discs - increasingly important for secondary storage e.g. CD, DVD

3.1 Disks


Organization of Disks(1)Organization of Disks(1) Components of Disk

Tracks

Sectors the smallest addressable portion of a disk

Disk pack a collection of a lot of platters

Cylinder (vertical collection of tracks) Arms are moving together

Seeking r/w arm movement the slowest part of reading data from disk

3.1 Disks


Organization of Disks(2)Organization of Disks(2)

Track Sector

Gaps

Surface of disk showing tracks and sectors

3.1 Disks


BoomRead/write heads

Spindle

Platters

Schematic illustration of disk drive

3.1 Disks

Organization of Disks(3)Organization of Disks(3)


Estimating Capacities(1)Estimating Capacities(1)

Cylinder --> Track --> Sector --> Byte Track capacity

= # of sectors per track * bytes per sector

Cylinder capacity = # of tracks per cylinder * track capacity

Drive capacity = # of cylinders * cylinder capacity

** Inner track, Outer track? Same capacity?

3.1 Disks


Estimating Capacities(2)Estimating Capacities(2)

Ex) file with 20,000 fixed-length data records on 300MB disk

# of bytes per sector = 512 # of sectors per track = 40 # of tracks per cylinder = 11 # of cylinders = 1,331

How many cylinders does the file require if each data record requires 256 bytes?

3.1 Disks


Organizing Tracks by Sector(1)Organizing Tracks by Sector(1) Two ways to organize data on a disk

by sector & by user-defined block

Physical placement of sectors logical view and physical view adequate way to view a file logically, but always not good way to

store sectors physically interleaving

— problems occur due to the gap between disk revolution speed

and disk controller speed— high performance disks (with high speed disk controller) now offer

1:1 interleaving

3.1 Disks


123

4

2122

23 323334

If interleaving factor is 3

..

..

.

Disk

3.1 Disks

• Interleaving

Organizing Tracks by Sector(2)Organizing Tracks by Sector(2)


Organizing Tracks by Sector(3)Organizing Tracks by Sector(3) Clusters (for improving file access speed)

One file consists of one or more clusters One cluster is a fixed number of contiguous sectors Cluster is the smallest unit of space Contiguous sectors per cluster and Mapping table No additional seek (arm movement) for one cluster Very useful when access granularity is medium large cluster vs. small cluster FAT (File Allocation Table)

contains a list of all the clusters in file each cluster entry --> physical location of the cluster

3.1 Disks


• FAT(File Allocation Table) FAT(File Allocation Table) - A table containing mappings to the physical locations of all the clusters in all files on disk storage

Cluster number

Clusterlocation

123...

File allocation table(FAT)

The part of the FAT pertaining to our file

.

.

.

2

1

3.1 Disks




Extents (further attempt for file access speed) Contiguous clusters per extent and Mapping table File may have several extents Very useful when accessing the whole file sequentially Many extents of a file increase amount of seeking

3.1 Disks

file extents extent

extent

extent

extent

extent extent



Fragmentation the space that goes unused within a cluster, block, track, or

other unit of physical storage internal fragmentation: loss of space within a sector or cluster trade-offs in use of large cluster sizes ex) size of sector is 512 bytes and size of all r

ecords in file is 300 bytes we may enforce one record / sector we may allow span sectors

3.1 Disks


Organizing Tracks by Block(1)Organizing Tracks by Block(1)

User-defined size of blocks can be fixed or variable No sector spanning ==> No internal fragmentation Blocking factor

the number of records that are to be stored in each block

In block-addressing, each block of data is accompanied by one or more subblocks containing extra information about the data block

Note: the “block” has a different meaning in the context of the UNIX I/O system

3.1 Disks


Sector organization vs.Sector organization vs. Block organization Block organization

1 1 1 1 1 1

1 1 1 1 1 1

1 1 1 2 2 2

2 2 3 3 3 4

4 4 4 5 5 5

Sector 1 Sector 2 Sector 3 Sector 4 Sector 5

(a) sector organization

(b) block organization

3.1 Disks

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2

2 2 3 3 3 4

4 4 4 5 5 5


Organizing Tracks by Block(2)Organizing Tracks by Block(2)

Subblocks count subblocks

contain the num of bytes in the accompanying data block key subblocks

contain key for last record in data block

Countsubblock Count subblock

Key subblock

Datasubblock

Countsubblock

Datasubblock

Countsubblock

Keysubblock

Datasubblock

Countsubblock

Keysubblock

Datasubblock

3.1 Disks


Nondata Overhead(1)Nondata Overhead(1)

Stored during preformatting Sector-addressable disks (physical)

sector/track address, condition, gaps, and synchronization marks

of no concern to programmer

Block-organized disks (user-defined) some of nondata for programmer more nondata provided with blocks than with sectors

3.1 Disks



Ex) block-addressable disks 20,000 bytes/track subblock, interblock gap equivalent to 300 bytes/block If we want to store a file containing 100-byte records, how

many records can be stored per track if blocking factor is 10 or 60 ?

if blocking factor is 10, then 20000 / ( (10*100)+300) = 15 blocks ==> 150 records

if blocking factor is 60, 180 records(=3 blocks)

3.1 Disks



Block-organized scheme larger blocking factor (many records in a block)

less amount of nondata block more efficient use of storage

pros. flexibility -> savings in time & efficiency

cons. internal fragmentation problem necessary for programmer and OS to do extra data

organization sector interleaving cannot be used to improve performance

3.1 Disks


Cost of Disk AccessCost of Disk Access Disk Access Time

Seek time + Rotational delay + Transfer time Seek time

time to move arm to locate cylinder more costly in multi-user environment average seek time for a particular file operation

Rotational delay time taken for disk to rotate correct sector under read/write

arm Transfer time (for actual data)

number of bytes transferrednumber of bytes on a track rotation time

3.1 Disks


Timing ComputationTiming Computation Seagate Cheetah 9 Gbyte drive

Minimum seek (track-to-track): 0.78msec Maximum seek: 19 msec Average seek: 8 msec Rotation delay: 3 msec Read one track: 6 msec Read one cluster: 0.28 msec

Read 8704K size file sequentially access the whole file: 1.7 sec randomly access the whole file: 9.25 sec


Effect of Block Size on Performance (In UNIX) Effect of Block Size on Performance (In UNIX)

UC Berkeley CSRG team Many experiments in BSD UNIX System Trade-off between block size, fragmentation, access time Doubling block size to 1,024 bytes improves performance by mo

re than a factor of 2 In 512-byte block size:

the slow throughput & low internal fragmentation (6.9%) In 4,096-byte block:

the fastest throughput & internal fragmentation (45.6%)

==> Invent the cluster concept!

3.1 Disks


Disk as BottleneckDisk as Bottleneck

Disk performance is increasing, but still slow! Even high-performance network is faster than disk

(5 M/ sec HD is slower than 100M bps LAN) Disk bound jobs

CPU and network must wait

Solution techniques multiprogramming striping - parallel I/Os avoid accessing disk

RAM disk, disk caches - locality

3.1 Disks


Magnetic TapeMagnetic Tape

A set of parallel tracks 9 tracks - parity bit Frame

one-bit-wide slice of tape

Interblock gaps permit stopping and starting

3.2 Magnetic Tape


Nine-track tapeNine-track tapeTrack Frame

011010010

Gap Data block Gap

3.2 Magnetic Tape


Estimating Tape Length(1)Estimating Tape Length(1)

There is an interblock gap for each data block Space requirement s s = n * ( b + g )

b is the physical length of a data block g is the length of an interblock gap n is the number of data blocks

Tape density Tape speed Size of interblock gap

3.2 Magnetic Tape


Estimating Tape Estimating Tape Length(2)Length(2)

Ex) one million 100-byte records 6,250 BPI tape 0.3 inches of interblock gap

How much tape is needed? when blocking factor is between 1 and 50

Nominal recording density Effective recording density:

num of byte per block / num of inches for block

3.2 Magnetic Tape


Estimating Data Transmission TimesEstimating Data Transmission Times

Factors of data transmission rate interblock gaps effective recording density nominal recording density speed of r/w head time to start/stop the tape

3.2 Magnetic Tape


Disk VS. TapeDisk VS. Tape

Disk Random access Immediate access Expensive seek in

sequential processing

Tape Sequential access Long-term storage No seek in sequential

processing

Decrease in cost of disk and RAM More RAM space is available in I/O buffers, so disk I/O decreases Tertiary storage for backup: CD-ROM, tape ...



Introduction to CD-ROMIntroduction to CD-ROM


CD-ROM: Compact Disc Read-Only Memory Can hold over 600MB(200,000 pages) Easy to replicate Useful for publishing or distributing medium But, not storing and retrieving data

CD-ROM is a child of CD audio CD audio provides

High storage capacity Moderate data transfer rate But, against high seek performance

Poor seek performance


History of CD-ROM: VideodiscHistory of CD-ROM: Videodisc

Videodisc technology developed in late 1960's and early 1970's The goal was to store movie A number of methods for storing video signals

1. Use a needle to respond mechanically to grooves in a disc like a vinyl LP record

2. Use optical storage Many companies were fighting which approach should become a

standard VideodiskLaserVisionCD audioCD-ROM



History of CD-ROM: LaserVisionHistory of CD-ROM: LaserVision

LaserVision Emerged as the winner of standard battles about standard Support CLV(Constant Linear Velocity) and CAV(Constant Angular

Velocity) format Fast seek performance by using CAV format Data are stored in analog form Why did they fail?

Earlier disputes over the physical format of the video disc, many incompatible encoding scheme and error correction techniques were used by many firms

No standard format the market never grew



History of CD-ROM: CD-ROMHistory of CD-ROM: CD-ROM CD-ROM

Philips and Sony developed CD-ROM in 1984 in order to store music on a disc

Use a digital data format The development of CD-ROM as a licensing system results in widely

acceptance in the industry Promised to provide a standard physical format Any CD-ROM drive can read any sector which they want Computer applications store data in a file not in terms of sector, thus, file

system standard should be needed In late 1985, videodisc/digital data industry moved into CD-ROM industry In early summer of 1986, an official standard for organizing files was

worked out




Physical Organization of Master DiskPhysical Organization of Master Disk

Master Disc Formed by using the digital data, 0 or 1 Made of glass and coated that is changed by the laser beam

Two part of CD-ROM Pit

The areas that is hit by the laser beam Scatter the light

Land Smooth, unchanged areas between pits Reflect the light

laser beam

pit

land

light


Encoding Scheme of CD-ROMEncoding Scheme of CD-ROM Encoding scheme

The alternating pattern of high- and low-density reflected light is the signal 1 : transition from pit to land and back again 0s : the amount of time between transitions

Constraint The limits of the resolution of the optical pickup (generic hardware limit),

there must be at least two 0’s between any pair of 1’s (no two adjacent 1s)

We cannot represent all bit patterns, thus, we need translation scheme We need at least 14 bits to represent 8 bits under this constraint

EFM(eight to fourteen modulation) coding is used

110 10000100000000EFM000000012



Format of CD-ROMFormat of CD-ROM CD audio chose CLV format instead of CAV format because

CD audio requires large storage space CD audio is played from the beginning to the end sequentially

Format of CD-ROM CLV(Constant Linear Velocity) A single spiral pattern Same amount of space for each sector Capability for writing all of sectors at the maximum density Rotational speed is slower in reading outer edge than in inner edge Finding the correct speed though trial and error Characteristics

Poor seek performance No straightforward way to jump to a specified location



Constant Angular Velocity DiskConstant Angular Velocity Disk

Magnetic disk usually uses CAV(Constant Angular Velocity) Concentric tracks and pie-shaped sectors Data density is higher in inner edge than in outer edge Storage waste: total storage is less than a half of CLV Spin the disc at the same speed for all positions Easy to find a specific location on a disk good seek

performance

CAVCLV



Addressing of CD-ROMAddressing of CD-ROM Addressing

Magnetic disk: cylinder/track/sector approach CD-ROM: a sector-addressing scheme

Track density varies thus, each second of playing time on a CD is divided into 75 sectors 75 sectors/sec, 2 Kbytes/sector At least one-hour of playing time Maximum capacity can be calculated: 600 Mbytes

60 min * 60 sec/min * 75 sectors/sec = 270,000 sectors

We address a given sector by referring minutes, second, and sector of play 16:22:34 means 34th sector in the 22nd second in the 16th minutes of play



Fundamental Design of the CD discFundamental Design of the CD disc

Initially designed for delivering digital audio information Store audio data in digital form Wave patterns should be converted into digital form Measure of the height of the sound: 65,536 different gradation(16 bits) Sampling rate: 44.1 kHz, because of 2 times of 20,000 Hz upto which

people can listen 16 bits sample, 44,100 times per second, and two channel for stereo

sound, we should store 176,400 bytes per seconds Storage capacity of CD is 75 sectors per seconds, we have 2,352 bytes

per sector CD-ROM divides this raw sector as shown in the following figure

12 bytessynch

4 bytessector ID

2,048 bytesuser data

4 byteserror

detection

8 bytesnull

276 byteserror

correction




File Structure Problem of CD-ROMFile Structure Problem of CD-ROM

The most important problem in using CD-ROM as a storage media Strongness and weakness of CD-ROM

Strong aspects of CD-ROM Data transfer rate: 75 sectors/sec Storage capacity : over 600 Mbytes Inexpensive to duplicate and durable

Weak aspects of CD-ROM Poor seek performance (weak random access)

Magnetic disk: 30 msec, CD-ROM : 500 msec Comparison of access time of a large file from several media

RAM: 20 sec, Disk: 58 days, CD-ROM: 2.5 years

We should have a good file structure avoiding seeks to an even greater extent that on magnetic disk


Storage as a HierarchyStorage as a Hierarchy

Primary storage 1/1,000,000,000 sec semi-conductors: registers, RAM, RAM disk, and disk cache

Secondary storage 1/1,000 sec magnetic disks(DASD), tape and mass storage(Serial)

Off-line storage (archival & back-up) 10 sec removable magnetic disks, optical discs, tapes



A Journey of a Byte(1)A Journey of a Byte(1)

What happens when a program writes a byte to a file on a disk?

WRITE(TEXT, c, 1); User Program ---> OS File manager -->

I/O buffer ---> I/O processor --->

Disk controller ---> Disk




1 The program asks the OS to write the contents of the variable c to the

next available position in TEXT.

2 The OS passes the job on to the file manager

3 The file manager looks up TEXT in a table containing information about it,

such as whether the file is open and available for use, what types of

access are allowed, if any, and what physical file the logical name TEXT

corresponds to.

4 The file manager searches a FAT for the physical location of the sector

that is to contain the byte.

5 The file manager makes sure that the last sector in the file has been

stored in a system I/O buffer in RAM, then deposits the ‘P’ into its proper

position in the buffer.

Logical layer




6 The file manager gives instructions to the I/O processor about where the

byte is stored in RAM and where it needs to be sent on the disk

7 The I/O processor finds a time when the drive is available to receive the

data and puts the data in proper format for the disk. It may also buffer the

data to send it out in chunks of the proper size for the disk

8 The I/O processor sends the data to the disk controller.

9 The controller instructs the drive to move the r/w head to the proper track,

waits for the desired sector to come under the r/w head, then sends the

byte to the drive to be deposited, bit-by-bit, on the surface of the disk.

Physical layer



What is buffer?What is buffer? Definition - the part of main memory available for storage of c

opies of disk blocks

Program buffers vs. System I/O buffers

Buffer manager subsystem responsible for the allocation for blocks goal:

minimize the number of disk access utilize the memory space effectively



Buffer ManagementBuffer Management

Buffer bottlenecks What if program performs both input and output on one

character at a time, and only one I/O buffer is available? At least two buffer - for input and output

I/O bound jobs: wait for I/O completion Buffering Strategies for Performance

use more than one buffer I/O system processes next block when CPU is processing

current one



Buffering Strategies(1): Buffering Strategies(1): Multiple bufferingMultiple buffering

Double buffering swapping roles of two buffers after I/O finished allows O/S operating on one buffer while the other buffer is

being loaded or emptied

any number of buffers can be used Buffer pooling

takes from a pool of available buffers decides which buffer to take from a buffer pool

take buffer by LRU



Double BufferingDouble Buffering

Program data area

Program data area

To disk

To diskI/O buffer 1

I/O buffer 1

I/O buffer 2

I/O buffer 2

(a)

(b)



Buffering Strategies(2): Buffering Strategies(2): Move & Locate modeMove & Locate mode

Move mode (using both system buffer & program buffer) moving data from one place in RAM to another before they

can be accessed sometimes, unnecessary data moves

Locate mode (using system buffer only or program buffer only) perform I/O directly between secondary storage and program

buffer (program’s data area) system buffers handle all I/Os, but program uses locations

through pointer variable



Move mode & Location modeMove mode & Location mode

system buffer

program’s data area

system buffer

disk

diskuser’s

programlocation(pointer)

Movemode

Locatemode



Buffering Strategies(3): Buffering Strategies(3): Scatter/Gather IOScatter/Gather IO

A file with many blocks and each block with header and data

Need to put the headers in one buffer and the data in a different buffer ==> may occur complication

Scatter-input mode a single READ can scatter data into a collection of buffers

Gather-output mode a single WRITE can gather several buffers and output



Scatter Input & Gather OutputScatter Input & Gather Output

disk

disk

scatterinput

gatheroutput

read()

write()

buffer 1

buffer 2

buffer 1

buffer 3

buffer 2



I/O in UNIXI/O in UNIX<Kernel I/O structure>

3.10 I/O in UNIX

I/O system

PROCESS

KERNEL

user programs libraries shell commands

system callinterface

block I/Osystem(normal files)

characterI/O system(terminals,printers, etc.)

networkI/O system(sockets)

block device drivers

disk disk...

character device drivers

consoles printers...

network interface drivers

...networks...


UNIX Kernel(1)UNIX Kernel(1) The central part of UNIX operating system View all I/O as operating on a sequence of bytes Make all operations below top layer independent of applic

ation’s logical view of file Data structures related to unix files

file-descriptor table: owned by user process open-file table: owned by kernel index-node: a kind of FAT (one inode for each file in use) index-node table: owned by kernel

3.10 I/O in UNIX


UNIX Kernel(2)UNIX Kernel(2)

File descriptor table associates each file descriptor to open file table every process has its own file descriptor table

Open file table entries for every open file file structures: r/w mode, offset, reference count

3.10 I/O in UNIX



file descriptor

file tableentry

o(keyboard)1(screen)2(error)3(normal)4(normal)5(normal)...

to open filetable

descriptor table

3.10 I/O in UNIX



R/Wmode

# ofprocessesusing it

Offsetof nextaccess

ptr towriteroutine

inodetableentry......

write 1 100 ......to inode table

write() routinefor this typeof file

open file table

3.10 I/O in UNIX



Inode(Index node) data structure used to describe a file when a file is opened, a copy of inode is loaded

into RAM for rapid access has a list of disk blocks of the file this list is UNIX counterpart to FAT

Device driver I/O processor program invoked by kernel

performing I/O for devices

3.10 I/O in UNIX


Inode StructureInode Structure

devicepermissions

owner’s useridfile size

block count

fileallocation

table

.

.

.

.

.

.

An inode

3.10 I/O in UNIX


Hard/Soft Link & File typesHard/Soft Link & File types

Linking file names to Linking file names to files hard link: direct reference

pointer from a directory to the inode of a file ln src-file target-file

soft(symbolic) link: pathnames link a file name to another file name ln -s src-file target-file

File typesFile types normal file (governed by block IO system) special file: stream, signal control of device (governed by ch

aracter IO system) sockets: endpoints of IPC (governed by network IO system)

3.10 I/O in UNIX


Let’s Review !!!Let’s Review !!!3.1 Disks

3.2 Magnetic Tape








3.10 I/O in UNIX

file structuresnu-oopsla lab1 chap3. secondary storage and system software 서울대학교...

Documents