Operating Systems.2 Flashcards
magnetic disks rotation rate
60-250 times/second
transfer rate
rate of data flow between drive and computer
positioning time/ random-access time
seek time + rotational latency
seek time
time taken to move head to desired cylinder (usually 3-12ms)
rotational latency
time for desired sector to rotate under disk head. based on spindle speed (60/RPM)
head crash
when disk head comes in contact with disk surface
connection between drive and computer
drive attached to computer using I/O bus- busses vary.
host controller in computer uses bus to talk to disk controller built into drive or storage array
storage range for hard disks
30GB - 3TB
access latency/average access time
average seek time + average latency
average I/O time
average access time + (amount to transfer/transfer rate) + controller overhead
SSDs advantages
- non volatile (used like a hard drive)
- more reliable than HDDs
- faster
- no moving parts so no seek time/rotational latency
SSDs disadvantages
- lower capacity
- expensive per MB
- last less longer
- busses too slow: can connect directly to PCI for example
how are disk drives addressed
large 1-dimensional arrays of logical blocks- logical block = smallest unit of data transfer
low level formatting
creates logical blocks on physical media. divides disk into sectors so disk controller can read and write
mapping of array onto disks
1D array of logical blocks mapped into sectors of the disk sequentially. first sector 0, then rest of sectors on the track, then rest of tracks in the cylinder then rest of cylinders from outermost to innermost
sector 0
first sector of first track on outermost cylinder
logical to physical address is easy unless
bad sectors. (non constant num. of sectors per track)
how are disks attached to system
an array of disks are attached and have controllers
disk scheduling
OS responsible for using hardware efficiently: fast access time + disk bandwidth
disk bandwidth
total bytes transferred/total time between first request and completion of last transfer
OS disks scheduling
maintains queue of requests per disk or device. disk is busy- request is added to queue as drive controller has a buffer. optimisation algorithms used when a queue exists. Several algorithms to service I/O operations in queue.
first come first serve (FCFS)
requests serviced in the order they arrive
(SSTF) shortest seek time first
request with minimum seek time from current head position is serviced next. form of SJF scheduling: can cause starvation of requests
SCAN aka elevator algorithm
start from head and go all the way to the other end and reverse till all requests are serviced.
C-SCAN
more uniform wait time than SCAN. Head goes all the way to the other end of the disk then returns straight to beginning of disk without servicing any requests on return trip- cylinders treated as a circular queue: wraps from last cylinder to first cylinder.
LOOK/C-LOOK
LOOK is to SCAN as C-LOOK is to C-SCAN. Difference is, instead of going to the very end of the disk, stop at last request in one direction before reversing/returning to beginning
what are the best algorithms for systems that place a heavy load on disk
SCAN, C-SCAN: less starvation
performance is dependent on
number and type of requests
reasonable choice for default algorithm
SSTF and LOOK
each sector holds what information
header info, data and ECC (usually 512 bytes but can be selectable)
to use a disk to hold files
OS needs to record its own data structures on disk- partition disks into cylinders- each treated as a logical disk. logical formatting: creating a file system
boot block
initialises system: loads bootstrap loader program stored in boot blocks of boot partition
swap space
Virtual memory uses (raw i.e. no file system) disk space as an extension to main memory
file
abstract data type: sequence of words/bytes: meaning is decided by programs/OS
file attributes
name: only information in human readable form
identifier: unique tag identifies file in system
type: needed for systems that supports different file types
size: current file size
location: pointer to where file is stored in system
protection: controls who can read/write/execute
time/date/user id: data for protection/security/usage monitoring
(info kept in directory structure)
file operations
create, delete, write/read (at pointer locations), seek, truncate (delete contents but keep attributes), open(Fi)/close(Fi)- search directory structure on disk for file entry Fi and move the content of entry to memory (+v.v.)
-must open to carry out any operation
how to manage open files
open file table: tracks open files
file pointer: points to last read/write location per process that has the file open
file count: counter of number of times a file is open- to allow removal of data when file is closed by the last process
file locking
at critical section: makes sure noone else has access to the file
- shared lock: reading file, many people can have access to file
- exclusive lock: writing a file
- mandatory: OS denies access depending on locks held and requested
- advisory: process can find status of locks and decide what to do
sequential-access file
when reading/writing: read/write one logical block at a time until output indicates end of file
to go backwards: must rewind all the way to the start
(read next/write next/ reset)
direct access
file is fixed length logical records. can read/write directly to relative block numbers (n)
read n/write n/position to n/read next/write next/rewrite n
how to implement sequential file access as direct access
(reset) cp = 0 read cp cp = cp + 1 write cp cp = cp + 1
file access methods: indexing
have an index of a file in memory for fast determination of location of the data to be operated on.
if too large: index in memory of the index on disk
directory structure
a collection of nodes containing information about all files- directory structure+files are both stored on disk
