Chapter 7 Device Management CIS 106 Microcomputer Operating

Chapter 7 Device Management CIS 106 Microcomputer Operating Systems Gina Rue CIS Faculty Ivy Tech State College Northwest Region 01

Introduction Device Management Device Manager: • Manages every peripheral device of the system • Maintains a balance of the user’s demand the system’s finite supply of devices See Illustration p. 145 2

Introduction Device Management • Four Basic Functions – tracking the status of each device, tape & disk drives, printers, terminals, etc. – uses policies to determine which process will get a device & for how long – allocating devices – deallocating devices • process level • job level 3

System Devices 3 Peripheral Device categories – dedicated • assigned only one job at a time, tape drives, printers, plotters – shared • assigned to several processes, disk pack, other storage devices by interleaving requests – virtual • combination dedicated & shared, printers, disks 4

Sequential Access Storage Media 2 Groups of Storage Media – Sequential access media • store records sequentially, one after another – Direct access storage devices • can store either sequentially or direct access, directly reading or writing to a specific place 5

Sequential Access Storage Media Magnetic Tape – developed for early systems for routine secondary storage, archiving, back-up data – records are stored serially, each record can be any length stored in any location, record length usually determined by application – time consuming to locate records, “fastforwarded” to desired position • parity bit used for routine error checking • interrecord gap (IRG) inserted in between each record See Fig. 7. 1 p. 147 6

Sequential Access Storage Media Magnetic Tape – blocking, grouping records into blocks before recording them, performed while file is created – transfer rate, density of tape (number or records in a block) – transport speed, speed of tape: transfer rate = density * transport speed 7

Sequential Access Storage Media Magnetic Tape Blocking Advantages: – fewer I/O operations – less tape is wasted Blocking Disadvantages – overhead & software routines are needed, deblocking, and record keeping – buffer space may be wasted Access times can vary widely 8

Direct Access Storage Devices – (DASDs) any devices that can directly read or write to a specific place on disk – Also called random access storage devices – Grouped into 2 categories • fixed read/write heads • movable read/write heads 9

Direct Access Storage Devices • Fixed-Head Drums & Disks – one of the first DASDs developed in early 1950’s – fixed-drum resembles a giant coffee can covered with magnetic film & formatted so tracks run around it – fixed-drum data is recorded serially on each track by the read/write head positioned over it See Fig. 7. 4 & 7. 5 p. 150 10

Direct Access Storage Devices • Fixed-Head Drums & Disks – fixed-disk looks like a phonograph album covered with magnetic film that has been formatted, into concentric circles (tracks) – fixed-disk data is recorded serially on each track by the read/write head positioned over it 11

Direct Access Storage Devices • Movable-Head Drums & Disks – Movable-Head drums have only a few read/write heads that move from track to cover the entire surface of the drum • least expensive device with only one read/write head for the entire drum • conventional design device with several read/write heads that move together See Fig. 7. 6 p. 151 12

Direct Access Storage Devices • Movable-Head Drums & Disks – Movable-Head disk have only one read/write head that floats over the surface of the disk • individual units, such as those with many PCs • disk packs, stack of disks, each disk has 2 surfaces for recording (top & bottom) • tracks on disk varies by manufacturer, but typically range from 200 to 800 • tracks are numbered, Track 0 is the outermost concentric circle See Fig. 7. 7 p. 151 13

Direct Access Storage Devices • Optical Disc Storage (ODS) – Includes CD-ROM, provides reliable highdensity storage for large amounts of data – 1995 ODS were read only, now CD-R (recordable) available – function similar to magnetic disk drive • read/write head on an arm • measure performance in data-transfer rate and average access time 14

Direct Access Storage Devices • Optical Disc Storage (ODS) – data transfer rate measured in kilobytes per second (Kps) refers to speed that data can be read off disc – access time indicates the average time required to move the read head to a specific place on the disc, expressed in milliseconds (ms) – hardware cache acts like a buffer by transferring blocks of data from the disk See Table 7. 2 p. 153 15

Direct Access Storage Devices Access Time Required: 3 factors that contribute to the time required to access a file • seek time • time required to position the read/write head on the proper track; slowest of 3 factors • search time • also known as rotational delay; time it takes to rotate the drum or disk until the requested record is moved under the read/write head • transfer time • when the data is actually transferred from secondary storage to main memory; fastest of 3 factors 16

Direct Access Storage Devices Access Time for Fixed-Head Devices – access a record by knowing its track and record number – total amount of time required to access data depends on 2 factors: • rotational speed • position of the record relative to the position of the read/write head Access Time = Search Time + Transfer Time See Fig 7. 8 p. 154 17

Direct Access Storage Devices Access Time for Movable-Head Devices – adds a third time element to the computation of access time: Access Time = Seek Time + Search Time + Transfer – overall, moveable-head devices are much more common than fixed-head DASDs • less costly and larger capacity; even though retrieval time is longer 18

I/O System Components • I/O subsystems work together – I/O Channels (dispatcher) • keeps up with the I/O requests from the CPU & pass them on to the appropriate control unit • programmable units placed between the CPU and the control unit • sends one signal for each function • synchronizes the fast speed of the CPU with the slow speed of the I/O device See Fig. 7. 9 p. 157 19

I/O System Components • I/O subsystems work together – I/O Control Unit (driver) • interprets the channel signal • sometimes part of the device • each control unit can direct several devices • flexibility, connect more than one channel for a control unit or connect more than one control unit to a single device • multiple paths increase reliability of the I/O subsystem See Fig. 7. 10 & 11 p. 158 -159 20

Communication Among Devices Demanding conditions of a busy computer system – Device Manager (3 Problems) • needs to know which components are busy & which are free • must be able to handle requests that come in during heavy I/O traffic • must accommodate disparity of speeds between CPU & I/O devices 21

Communication Among Devices Concurrent processing & I/O • each unit in the I/O subsystem can finish its operation independently • after a device has begun writing a record, & before completing the task, the connection between the device & its controller can be cut off to begin another I/O task with another device • CPU is free to process data while I/O is being performed 22

Communication Among Devices How does the system know when a device has completed an operation? • Hardware flag must be tested by the CPU • made up of 3 bits & resides in the Channel Status Word (CSW) – CSW, predefined location in main memory, contains channel status information • each bit represents one of the components indicating channel status – channel bit, control unit, & device – 0 free, 1 busy 23

Communication Among Devices • Each component has access to the flag • can be tested before proceeding with the next I/O operation – Common ways to perform this test: • polling • interrupts 24

Communication Among Devices • Polling – use special machine instructions to test the flag – major disadvantage CPU wastes time testing flag • Interrupts – more efficient way to test flags – interrupt handler determines best course for action 25

Communication Among Devices • Direct Memory Access (DMA) – allows a control unit to access main memory more directly – data can be transferred to & from memory without CPU interaction • Buffers – more efficient way to test flags – interrupt handler determines best course for action See Fig. 7. 12 p. 161 26

Management of I/O Requests • Device Manager divides the task into 3 parts handled by a software component of the I/O subsystem • I/O traffic controller • I/O scheduler • I/O device handler 27

Management of I/O Requests • I/ traffic controller – monitors every device status, control unit, & channel – maintains control blocks – three main tasks: – determines if path is available – determines best path – determines when a path will become available See Table 7. 5 p. 162 28

Management of I/O Requests • I/O scheduler • performs same job as the Process Scheduler • allocates the devices, control units, & channels • decides which request will take priority • synchronizes its work with the traffic controller 29

Management of I/O Requests • I/O Device handler • processes the I/O interrupts • handles error conditions • provides detailed scheduled algorithms • each type of I/O device has its own device handler algorithm 30

Device Handler Seek Strategies • Seek Strategy determines the order processes get the device, keeping seek time to a minimum • Every scheduling algorithm should do the following: • minimize movement • minimize mean response time • minimize the variance in response time 31

Device Handler Seek Strategies • First come first served (FCFS) – simplest device-scheduling algorithm – easy to users & fair to users – seek time delay disadvantage – seek time is the most time consuming See Fig. 7. 13 p. 164 32

Device Handler Seek Strategies • Shortest seek time first (SSTF) – shortest jobs are processed first – longer jobs are made to wait – requests the track closest to the one being served – favors easy-to-reach requests & postpones long jobs farther away See Fig. 7. 14 p. 164 33

Device Handler Seek Strategies • SCAN uses a directional bit to indicate arm movement • LOOK (elevator algorithm/SCAN variation), arm does not go all the way to either edge unless requested • N-step SCAN holds all requests until the arm starts on its way back See Fig. 7. 15 p. 165 34

Device Handler Seek Strategies • C-SCAN (Circular SCAN) the arm picks up requests on its path during inward sweep • C-LOOK arm does not move all the way back to last track unless required The best scheduling algorithm may be a combination of more than one scheme 35

Search Strategies: Rotational Ordering Optimize search times by ordering requests once the read/write heads have been positioned 36

Search Strategies: Rotational Ordering • “Rotational Ordering” – amount of time wasted due to rotational delay can be reduced – requests are ordered within each track – device controller must provide “rotational sensing” so the device driver can “see” which sector is under the read/write head See Fig. 7. 16 & table 7. 6 p. 167 37

Summary • Device Manager’s job to manage all the devices despite varying speeds & sharing abilities • Direct Access and Sequential Access • One or many read/write heads • Fixed position or heads move across surface 38

Summary • Complex task of balancing demand for devices divided among hardware components: – channels – control units – devices • Success of the I/O subsystem depends on the communications that link these parts 39

Summary • Seek strategies to optimize seek time – FCFS: First Come, First Served – SSTF: Shortest Seek Time First – SCAN/LOOK – N-step SCAN – C-SCAN/C-LOOK 40