ICS 431-Ch11-Mass Storage Structure

Welcome to the ICS 431 sessions and let us
collaborate to understand and to be knowledgeable in the

“Operating Systems”
Dr. Tarek Helmy, KFUPM-ICS

Operating Systems ICS 431
Weeks 12-13
Ch11: Mass-Storage Structure
Dr. Tarek Helmy El-Basuny

Dr. Tarek Helmy, KFUPM-ICS 2
Mass-Storage Structure
• Memory System Hierarchy
• The computer system combines many storage technologies to balance the cost and the benefits
• Overview of Mass-Storage Structure
– The mass-storage formatted to: Tracks, sectors, blocks, cylinders,
• Disk Performance and Parameters
– Access time = Seek time + Latency time + Data transfer time.
– Which parameters are static and which are dynamic?
• Disk’s Head scheduling will try to improve both access time and transfer rate by scheduling the
requests of the disk I/Os in a good order.
• Disk Scheduling Algorithms
– FIFO, SSTF, Scan (The Elevator Algorithm), C-Scan, C-Lock.
• Selecting a Disk-Scheduling Algorithm
• Disk Formatting: Physical (Low Level) and Logical (High level)
• Disk Partitioning: Can we minimize the seek time through partitioning?
• Important Aspects of Mass Storage management: Reliability, Availability and Integrity.
– Reliability: Can we detect errors and recover them? To improve the reliability
– Availability: Is the disks always available to the user if one of them fails?
– Integrity: we must know exactly what is lost when something goes wrong
• Enhances in disks technology motivated HW engineers to attach many disks to a computer
system and allow them to act in parallel as a single logical disk (Redundant Array of
Independent Disks, RAID) so that we can improve: Reliability, Availability and Integrity.
• Data Striping, and RAID Levels (0~6). 3
Memory System Hierarchy
• Combining many technologies to balance costs/efficiency.
CPU registers hold words
L0: retrieved from L1 cache.
Registers
SRAM VS. DRAM
Smaller in size,
L1 cache holds cache • SRAM is on-chip while DRAM
Faster Access,
L1: On-chip L1 lines retrieved from the L2 is off-chip.
and
cache (SRAM) cache memory. • SRAM is faster compared to
Expensive
DRAM.
L2 cache holds cache
• SRAM consumes less power
L2: Off-chip L2 lines retrieved from
than DRAM because DRAM
cache (SRAM) main memory.
needs to be refreshed.
• SRAM uses more transistors
per bit of memory compared
Main memory holds
L3: Main memory to DRAM.
disk blocks retrieved
(DRAM) from local disks. • DRAM is available in larger
storage capacity while SRAM
is of smaller size.
Larger in size, Local disks hold • SRAM is more expensive than
Slower access, files retrieved DRAM.
and Local secondary storage from disks on • Cheaper DRAM is used in
L4:
Cheaper remote network main memory while SRAM is
(Local Disks: EMD or SSD) servers. commonly used in cache
memory.
Remote
Remote secondary storage Storage on
L5:
(Distributed File Systems, Web Servers) network
servers.
4
Sold-State Disks
• SSD stores persistent data, it is a type of DRAM or it may be

called Non-volatile Memory Devices (NVMD).
• SSD devices are more reliable than HDDs because they have no
moving parts and faster because they have no seek or rotational
latency times.
• Reliable in portable environments,
• No noise as it has no moving parts,
• Faster start up,
– Does not need spin up
• Extremely low read latency,
– No seek time (25 us per page/4KB)
• Deterministic read performance,
– The performance does not depend on the location of data.
• Costs more, depending on the unit capacity.
5
What is inside a Hard Disk Drive?
• The Hard Disk contains platters, each with two surfaces.

• Each surface organized into concentric rings called tracks.
• Each track consists of sectors separated by gaps.
Spindle The read/write head

Arm is attached to the end
Platters of the arm and flies over
the disk surface.
Actuator
Tracks
surface
Spindle
R/W Head
SCSI
Sector
connector
6
Hard Disk Structure
 Tracks: Multiple concentric circles
containing data.
 Sectors: Subsections of a track.
 Blocks: Subsection of a sector created
during formatting (512-4096 B).
• A: Track,
 Cylinder: All tracks of same diameter in • C: Track Sector
the disk pack. • B: Sector,
Disk Data Layout • D: Cluster
 Moveable heads: One r/w head per
surface.
 Seek time: Time to get r/w head to the
desired track.
 Head crash: Results from disk head
making contact with the disk surface.
 Rotation/latency delay: Waiting time for
the desired data to spin under R/W
head.
 Block transfer time: Time to read or
write a block of data.
7
Disk Performance Parameters
 To read or write, the head must be positioned at the desired track and
at the beginning of the desired sector.
 Seek time: Time it takes to position the head at the desired track.
 Rotational delay/latency: The time it takes for the disk to rotate so that
the required sector is above the disk’s head .
– Drives rotate at 60 to 200 times per second.
• Transfer rate is the rate at which data flows between the drive and the
MM. Data transfer occurs as the sector moves under the head.
 Access time = Seek time + Latency time (about 15-60ms) + Data
transfer time (about1-2ms).
 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).
8
Timing of a Disk I/O Transfer
Total Average Access Time: Ta = Ts + 1/(2*r) + b / (r*N)

Where: Ts= Average seek time
b = Number of bytes to be transferred
N = Number of bytes on a track
r = Rotation speed in revolutions per second
• Transfer Rate to/from disk depends on the rotation speed of the disk and the
track location, so it will be higher for data on the outer tracks (where there
are more data sectors) and lower towards the inner tracks (where there are
fewer data sectors).
• In general the Transfer Rate depends on the # of bytes on track and rotation
speed. 9
Disk’s Head Scheduling
 The OS is responsible for using the mass-storage in an efficient manner.
 This means having a fast access time and large disk bandwidth.
 The goal is to minimize the seek time (Seek time  seek distance)
 The idea is to improve both access time and bandwidth by scheduling the
requests of the disk I/Os in a good order.
 Whenever a process needs an I/O to or from the disk, it issues a system call to
the OS, where the request specifies:
 Whether this operation is input or output.
 What is the disk address to read from?
 What is the memory address to write into?
 What is the number of bytes to be transferred?
 If the desired disk drive and the controller are available, the request can be
serviced immediately, otherwise the request will be queued for that drive.
10
Disk’s Head Scheduling Algorithms
• Disk’s Head Scheduler: Examines the current contents of the disk’s queue and
decides which request to be served next.
• The head scheduler must consider 3 important factors:
1. Overall performance of the disk
2. Fairness in treating processes’ disk requests
3. Cost of executing scheduling algorithm
 Several algorithms exist to schedule the disk I/O requests.
 First-In First-Out (FIFO),
 Shortest Seek Time First (SSTF)
 SCAN: The Elevator Algorithm,
 C-SCAN, and C-LOOK
• These algorithms will be used with electromechanical disks (HDD) where they have
spinning disks and movable heads.
• They will not be used with solid state disks or electronic disks.
• These algorithms will be evaluated based on the head movement distance that will
affect the seek time. We will illustrate that with the following example:
• Disk queue (0-199) cylinders with requests for I/O to blocks on cylinders 98, 183, 37,
122, 14, 124, 65, 67, with head at cylinder 53. 11
First-In First-Out (FIFO)
– Requests will be processed sequentially according to the order.

– Fair to all processes.
Head at 53 98 183 37 122 14 124 65 67
0 14 37 53 65 67 98 122 124 183 199
Traveled Distance
45
45
85
+ 85
146
+ 146
85
+ 85
108
+ 108
110
+ 110
59
+ 59
2 + 2
= 640
12
SSTF – Shortest Seek Time First
 When a disk operation finishes, the OS chooses the closest request to

the current head position to be processed (the one that minimizes seek
time).
 Selects the request with the minimum seek time from the current head
position.
 SSTF scheduling is a form of SJF scheduling.
 This algorithm minimizes latency and thus gives the best overall
performance.
 It suffers from poor fairness. Requests will get varying responses
depending on how lucky they are in being close to the current location of
the heads.
 In the worst case, some requests may starve (long delayed).
13
SSTF – Shortest Seek Time First
– Requests will be processed based on the closest first.

– Some request will starve
Head at 53 98 183 37 122 14 124 65 67
0 14 37 53 65 67 98 122 124 183 199
Traveled Distance
12
12
2
+ 2
30
+ 30
23
+ 23
84
+ 84
24
+ 24
2 + 2
59
+ 59
= 236
SCAN: The Elevator Algorithm
• The disk head moves (from the current position to the center of the disk) serving the
closest requests in that direction. When it runs out of requests in the direction it is
currently moving, it switches to the opposite (to the edge) direction doing the same.
Head at 53 98 183 37 122 14 124 65 67

Center Edge
0 14 37 53 65 67 98 122 124 183 199
Traveled Distance
16
16
+ 23
23
+ 14
14
+ 65
65
+ 2
2 + 31
31
+ 24
24
+ 2
2 + 59
59
= 236
Some Modification to the Scan
• Reverse the direction immediately, without going all the way to the edge or to
the center of the disk.
Head at 53 98 183 37 122 14 124 65 67
0 14 37 53 65 67 98 122 124 183 199
Traveled Distance
16
16
23
+ 23
51
+ 51
2 + 2
31
+ 31
24
+ 24
2 + 2
59
+ 59
= 208
SCAN: The Elevator Algorithm
 This algorithm usually gives fair service to all requests, but in the worst case, it
can still lead to starvation.
 While it is satisfying requests on one cylinder, other requests for the same
cylinder could arrive.
 If enough requests for the same cylinder keep coming, the heads would stay at
that cylinder forever, starving all other requests.
 This problem is easily avoided by limiting how long the heads will stay at any
cylinder. One simple scheme is to serve only the requests for the cylinder that
are already there when the head gets there.
 New requests for that cylinder that arrive while existing requests are being
served will have to wait for the next pass.
17
Circular-SCAN (C-SCAN or LOOK)
 Circular scanning works just like the elevator

to some extent.
 Provides a more uniform wait time than
SCAN.
 The head moves from the current position
toward the nearest end or the one that The head moves towards the edge
maximizes the rate, servicing requests as it
goes.
 When it reaches the other end, however, it
jumps to the other end without servicing any Fig. 1
requests on the return trip.
 Once it hits the bottom or top it jumps to,
moves in the same direction servicing
requests as it goes.
 Keep in mind that the huge jump doesn't
count as a head movement
 Treats the cylinders as a circular list that The head moves towards the center
wraps around from the last cylinder to the
The total head movement for this algorithm is
first one. only 187 track
 The total head movement of Fig. 1 is ??? cylinders. Please calculate the head movements.
18
C-LOOK
 It is a modified version of C-SCAN

 Arm only goes as far as the last
request in each direction.
 Then reverses direction
immediately, without going all the
way to the end or to the center of
the disk.
 Versions of SCAN and C-SCAN

that follow this pattern are called
LOOK and C-LOOK scheduling,
because they look for a request
before continuing to move in a
given direction.
 Illustration shows total head movement of ??? cylinders. Please

calculate the head movements. About 30 cylinders less than the C-Scan.
19
Selecting a Disk-Scheduling Algorithm
 What is the best disk’s head scheduling algorithm?
 The answer depends on too many factors, i.e. types of file, number of requests, allocation
method, etc.
 SSTF is common and has a natural appeal if the process needs high data transfer rate.
 SCAN and C-SCAN perform better for systems that place a heavy load on the disk.
 Suppose that the queue has just one or two requests, then all algorithms behave the
same as FCFS.
 Selecting the best algorithm can be biased by the file-allocation method, contiguous,
linked, indexed.
 In general, either SSTF or C-LOOK is a reasonable choice for the default/normal
processes.
 Recommended actions:
 The disk-scheduling algorithm should be written as a separate module of the OS,
allowing it to be replaced with a different algorithm if necessary or to be updated.
 The location of directories and index blocks is also important. Opening a file needs
to search the directory on the disk and then reading the file needs to access the disk
again. Caching the directory and disk blocks in main memory can also help in reducing
the head movement, specially for reading operation. 20
Disk Formatting
• Two formatting processes are required before we can write data to a HDD:
1. Low-level (Physical) formatting: usually performed at the factory
 Marking out cylinders and tracks for a blank hard disk, and then
dividing tracks into multiple sectors.
 Sequentially numbers the tracks and sectors on the disk,
 Identifies each track and sector,
 Disk is physically prepared to hold data.
2. High-level (Logical) formatting: usually performed by the OS.

 Determines how the OS uses a disk and labels logical bad sectors.
 Builds data structure to keep track of directories and files (initializes
FAT, and directory structure tables)
 Builds the data structures used by the OS to identify the logical drive
or partition's contents .
 Sets the bootable partition.
21
Disk Partitioning and MBR
• Creating partitions on a hard drive enables it to support separate file systems,

each in its own partition. In fact, each partition is like a separate disk.
• Master Boot Record (MBR) is the first sector of a hard drive (cylinder 0, track 0,
sector 1):
– MBR Identifies how and where an OS is located so that it can be loaded into
the computer's MM.
• What happens while turning on the computer?

– BIOS executed and loads the MBR.
– MBR also contains information about the HD’s partitions.
– There is a flag to identify the boot partition.
– MBR then loads OS kernel into the MM.
22
Disk Partitioning and the Seek Time
• If seek distances can be minimized then the overall performance will be improved and the
importance of the disk-head scheduling algorithm will be narrowed.
• Disks are typically partitioned to minimize the largest possible seek time.
– A partition is a collection of cylinders
– Each partition is logically a separate disk
• Swap-space (virtual memory): The disk space that is used as an extension to the main
memory.
• Swap-space can be stamped out of the normal file system or, more commonly, it can be in
a separate disk partition.
• Knowing that the transfer rate depends on the position of reading or writing, do you think
virtual memory should be selected from outer or inner partitions?
Partition A Partition B
23
Mass Storage Performance
 Three important aspects of Mass Storage management:

– Reliability: Can we retrieve the information in case of a disk failure.
• Storing extra information that is not needed normally but that can
be used in case of failure of a disk. i.e.
• Redundancy is the main hack to increase the reliability.
• Error Correcting Code (ECC) and various correction schemes
improve the reliability (but cannot improve Availability).
– Availability: Is the system always available to the user if something
fails? Availability is improved by adding redundant disks.
• If a single disk fails, the required information can be accessed from
the redundant disk.
• Capacity penalty to store redundant information.
• Bandwidth penalty to update both disks.
– Data Integrity
• Can we know exactly what is lost when something goes wrong.
24
Handling Bad Sectors on the HDD
A. To improve both reliability and availability of the HDD, the OS

reserves spare sectors to:
– Substitute a spare for the bad sector (sector sparing) or
– Shift all sectors to bypass bad one (sector forwarding)
25
Reliability improve by using ECC (Parity Bit)
• A parity bit can be used to detect and correct a one-bit error ?

– Suppose you have a binary number, represented as a collection of
bits: <b3, b2, b1, b0>, e.g. 0110
• We can add one extra bit (parity) to reflect filliping of any bits.
• Odd and even parity:
– Count the number of bits that are (ones), see if it’s odd or even.
• EVEN parity: if the number of 1 bits is even set the parity bit to 0.
• ODD parity: if the number of 1 bits is odd set the parity bit to 1.
– Parity(<b3, b2, b1, b0 >) = P0 = b0  b1  b2  b3
– Parity(<b3, b2, b1, b0, P0>) = 0 if all bits are intact
– Parity(0110) = 0, Parity(01100) = 0
– Parity(11100) = 1 => ERROR!
– Parity can detect a single bit error, but can’t tell you which of the bits
got flipped.
26
Reliability improve by using ECC (Hamming Code)
• Hamming code can detect double bits error and detect & correct single bit
errors in a message. Hamming Code is composed on 3 bits as following:
– h0 = b0  b1  b3
– h1 = b0  b2  b3
– h2 = b1  b2  b3
– If the code is (1101)
– h0 (<1  0  1>) = 0
– h1 (<1  1  1>) = 1
– h2 (<1  1  0>) = 0
– The hamming code is <010>
– Message <1101>, with hamming = <b3, b2, b1, h2, b0, h1, h0> = <1100110>
– If a bit is flipped in the message, e.g. <1110110>
– Hamming of the error message (<1111>) = <h2, h1, h0> = <111>
– Now the difference between the two hamming codes <010>, <111> is <101>
– The Error has occurred in bit # 5.
27
Disks Technology Trends Price per Megabyte of DRAM
• Disks are getting smaller in size

– Smaller  spin faster; smaller distance for head
to travel; and lighter weight.
• Disks are getting denser
– More bits/square inch  small disks with large
capacities.
• Disks are getting faster
– Seek time, rotation latency: 5-10% faster/year
– Bandwidth: 20-30% larger/year
• Disks are getting cheaper
• Overall:
– Disk capacities are improving much faster than
performance.
• A fixed disk drive is likely to be more reliable than a
removable disk.
• So it is now economically feasible to attach many
disks to a computer system and allow them to act in
parallel as a single logical disk.
• Having a large number of disks in a system
improves the rate of R/W and the reliability. Price per Megabyte of Magnetic Hard Disk
28
RAID Technology
• Because disks are getting small and cheap, so it’s easy to put lots of
disks (10s to 100s) in one box to increase the storage, performance,
and availability.
• RAID is a Redundant Array of Inexpensive/Independent Disks

viewed by the OS as a Single Logical Expensive Disk (SLED).
 With parallel access to multiple disks, the OS can improve the transfer
rate.
• The RAID box with a RAID controller looks just like a SLED to the
computer.
• Redundant copies and Error Correction Code (ECC) are striped

across the disks to increase reliability and availability.
29
RAID Levels
• Redundancy can be implemented either by:

– Duplicating/mirroring: replicates data on more than one disk.
– Striping: is a technique that breaks up a file and spreads the data across
all the disk drives in a RAID group.
• Seven RAID levels (0~6) were introduced, because of:
– Cost
– Performance
– Fault-tolerance characteristics
– Different storage needs
• They were used in different environments for different applications:
– Banking systems,
– Storage systems,
– Responsive services,
30
Improvement of Disk Performance
 Data Striping:
– Bit-level striping: Strip the bits of each byte across multiple disks.
• No. of disks can be a multiple of 8 or divides of 8.
– Byte-level stripping: Strip the bytes across multiple disks.
– Block-level striping: Blocks of a file are striped across multiple disks;
with n disks, block i goes to disk (i mod n)+1
• Every disk participates in every access
– Number of I/O per second is the same as a single disk
– Number of data read/written per second is improved
• Provide high data-transfer rates.
– Throughput is increased through a larger effective block size and
through the ability to perform parallel I/O.
31
Disk Striping (RAID-0)
• Disk striping (RAID-0)

– Disk blocks are striped and stored on separate disks.
– Provides for higher disk bandwidth through a larger effective block size
• If the rotation of the disks can be synchronized so that all have the same
rotational latency for any read/write operation then we have what is known as
RAID.
1 2 3
OS disk
block
8 9 10 11
12 13 14 15 8 9 10 11 12 13 14 15 0 1 2 3
0 1 2 3
Physical disk blocks 32

Data Mapping for RAID Level 0
• In RAID Level 0 files are striped Logical Disk
across disks, no redundant info which

makes the reliability lower than with strip 0
single disk (if any disk in the array strip 1
strip 0 strip 1 strip 2 strip 3
fails, we have a problem)
strip 2
• High read throughput/data transfer strip 3
strip 4 strip 5 strip 6 strip 7
strip 9 strip 10 strip 11
rate strip 4
strip 8
strip 13 strip 15
 Strips 0, 1, 2, and 3 can be strip 12 strip 14
strip 5
accessed in parallel
strip 6
• Best write throughput (no redundant
info to write) strip 7
strip 8
 High I/O request rate strip 9
Array
Management
 Servicing multiple I/O requests in strip 10 Software
parallel, e.g., requests for strip 0 strip11
and strip 3 can be serviced strip 12
simultaneously. strip 13
• Makes disks share the load, strip 14

strip 15
improving Latency: less queuing
delay – a queue for each disk.
33
RAID Level 1
Data Disks Mirror Copies
 Redundancy is achieved by using mirroring (duplication of data on other disks)

where each disk is fully copied on its “shadow”.
 Files written to both, if one fails, flags it and gets the data from the mirror.
 Expensive due to the need of mirror disks.
 A read can be serviced by either disks (the least access time).
 A write requires updating two strips, Logical write = two physical writes.
 Recovery is simple.
34
RAID Level 2
Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6
Data Disks ECC Disks
 Bit-level striping with (Hamming ECC) for error correction & recovery.
 All member disks participate in the execution of every I/O request.
(Gives high transfer rate but not I/O request rate).
 Spindles and heads are all synchronized to the same position.
 Requires smaller number of disks compared to RAID level 1.
 It uses parity checks, associate a parity bit with each byte in the
memory, even parity or odd parity.
 In RAID level 2, 3 parity bits are used to reconstruct the damaged byte.
35
Raid Level 3
• Striping at the byte level and stores dedicated parity bits on a separate disk
drive
• Because RAID 3 combines parity and striping with stored parity bits on a
dedicated disk,
• The disks must spin in sync, so sequential read/write (R/W) operations
achieve good performance.
• A read accesses all the disks.
• A write accesses all disks plus the parity disk.
• On a disk failure, read data from the remaining disks plus parity disk to
compute the missing data.
• High throughput for transferring large amounts of data
Byte 0 Byte 1 Byte 2 Byte 3
Parity Single parity disk can be used

to detect and correct errors.
Parity disk
Data Disks
36
RAID Level 3: Parity Disk
10010011
11001101 P
10010011
...
logical record 1 1 1 1
0 1 0 1
Striped physical 0 0 0 0
records 1 0 1 0
0 1 0 1
0 1 0 1
1 0 1 1
1 1 1 1
25% capacity cost for parity in this configuration (1/N)
37
Raid Level 4
• Block-interleaved parity
– RAID 4 is very similar to RAID 3. The main difference is the way of sharing
data. They are divided into blocks and written on disks.
– One disk is a parity disk, keeps parity blocks
– Parity block at position X is the parity for all blocks whose position is X on
any of the data disks
– A read accesses only the data disk where the data is there.
– A write must update the data block and its parity block.
– Can recover from an error on only one disk.
– Note that with N disks we have N-1 data disks and only one parity disk, but
can still recover when one disk fails
– But write performance worse than with one disk (all writes must read and
then write the parity disk)
Stripe 0 Stripe 1 Stripe 2 Stripe 3 P0-3

Parity Disk 38
Data Disks
Raid Level 5
– Like RAID 4, but parity blocks distributed to all disks

– Read accesses only the data disk where the data is
– A write must update the data block and its parity block
• But now all disks share the parity update load
Stripe Stripe Stripe Stripe P0-3

0 1 2 3
Stripe Stripe Strip P4- Stripe

4 5 e6 7 7
Stripe P8- Stripe Stripe

Stripe
8 11 10 11
9
Data and Parity Disks
39
Raid Level 6
• Level 5 with an extra parity bit
• Two different (P and Q) check blocks
– Each protection group has
• N-2 data blocks
• One parity block
• Another check block (not the same as parity)
• Can recover when two disks are lost
– Think of P as the sum and Q as the product of D blocks
– If two blocks are missing, solve equations to get both back
• More space overhead (only N-2 of N are data)
• More write overhead (must update both P and Q)
– P and Q still distributed like in RAID 5
Stripe 0 Stripe 1 Stripe 2 Stripe 3 P0-3 Q0-3
Parity Disks
Data Disks 40
Selecting a RAID Level
• If a disk fails, the time to rebuild its data can be significant and will
vary with the RAID level used.
• RAID level 0 is used in high-performance applications where data
loss is not critical.
• Rebuilding is easiest for RAID level 1. Simply copy the data from
another disk.
• RAID level 1 is popular for applications that require high
reliability with fast recovery.
• The combination of RAID levels 0 and 1 (RAID 0 + 1) is used for
applications where performance and reliability are very important, for
e.g. banking system’s databases.
• Due to RAID 1’s high space overhead, RAID level 5 is often preferred
for storing large volumes of data.
• RAID level 6 is not supported currently by many implementations, but
should offer better reliability than level 5.
41
The End!!
Thank you
Any Questions?
42

ICS 431-Ch11-Mass Storage Structure

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ICS 431-Ch11-Mass Storage Structure

Uploaded by

Copyright:

Available Formats

Welcome to the ICS 431 sessions and let us

collaborate to understand and to be knowledgeable in the

Dr. Tarek Helmy, KFUPM-ICS

Ch11: Mass-Storage Structure

Dr. Tarek Helmy El-Basuny

• SSD stores persistent data, it is a type of DRAM or it may be

• The Hard Disk contains platters, each with two surfaces.

Spindle The read/write head

Total Average Access Time: Ta = Ts + 1/(2*r) + b / (r*N)

 The OS is responsible for using the mass-storage in an efficient manner.

– Requests will be processed sequentially according to the order.

0 14 37 53 65 67 98 122 124 183 199

 When a disk operation finishes, the OS chooses the closest request to

– Requests will be processed based on the closest first.

0 14 37 53 65 67 98 122 124 183 199

Head at 53 98 183 37 122 14 124 65 67

Head at 53 98 183 37 122 14 124 65 67

0 14 37 53 65 67 98 122 124 183 199

 Circular scanning works just like the elevator

 It is a modified version of C-SCAN

 Versions of SCAN and C-SCAN

 Illustration shows total head movement of ??? cylinders. Please

2. High-level (Logical) formatting: usually performed by the OS.

• Creating partitions on a hard drive enables it to support separate file systems,

• What happens while turning on the computer?

 Three important aspects of Mass Storage management:

A. To improve both reliability and availability of the HDD, the OS

• A parity bit can be used to detect and correct a one-bit error ?

• Disks are getting smaller in size

• RAID is a Redundant Array of Inexpensive/Independent Disks

• Redundant copies and Error Correction Code (ECC) are striped

• Redundancy can be implemented either by:

• Disk striping (RAID-0)

Physical disk blocks 32

• In RAID Level 0 files are striped Logical Disk

across disks, no redundant info which

• Makes disks share the load, strip 14

Data Disks Mirror Copies

 Redundancy is achieved by using mirroring (duplication of data on other disks)

Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6

Data Disks ECC Disks

Byte 0 Byte 1 Byte 2 Byte 3

Parity Single parity disk can be used

25% capacity cost for parity in this configuration (1/N)

Stripe 0 Stripe 1 Stripe 2 Stripe 3 P0-3

Stripe 4 Stripe 5 Stripe 6 Stripe 7 P4-7

– Like RAID 4, but parity blocks distributed to all disks

Stripe Stripe Stripe Stripe P0-3

Stripe Stripe Strip P4- Stripe

Stripe P8- Stripe Stripe

Data and Parity Disks

Stripe 0 Stripe 1 Stripe 2 Stripe 3 P0-3 Q0-3

Stripe 4 Stripe 5 Stripe 6 Stripe 7 P4-7 Q4-7

Stripe 8 Stripe 9 Stripe 10 Stripe 11 P8-11 Q8-11

You might also like

Total Average Access Time: Ta = Ts + 1/(2r) + b / (rN)