Computer Organization and Technology External Memory Assoc. Prof. Dr. Wattanapong Kurdthongmee Division of Computer Engineering, School of Engineering and Resources, Walailak University 1
Magnetic Disk A disk: A circular platter constructed of nonmagnetic material (substrate); i.e. aluminum or aluminum alloy or glass (which has many benefits), Coated with a magnetizable material. Data are recorded/retrieved via a conducting coil: head During a read or write, the head is stationary while the platter rotates. Write mechanism: When electricity flows through a coil, it produces a magnetic field, Electric pulses are sent to the write head magnetic patterns are recorded on the platter s surface. Different patterns for positive and negative currents. Read mechanism: Traditionally, magnetic field moving relative to a coil produces an electrical current in the coil. 2
Magnetic Disk Track width A disk: Read mechanism: Currently, for higher frequency operation, the read head consists of a partially shielded magnetoresistive (MR) sensor. The MR sensor has an electrical resistance that depends on the direction of the magnetization of the medium moving under it. Write current Read current Shield Inductive write element Magnetization N S S N N S S N N S N S MR Sensor 3
Magnetic Disk Data organization and formatting: The data on the platter are organized in a concentric set of rings called tracks. Tracks: Each track is the same width as the head. There are thousands of tracks per surface. Adjacent tracks are separated by gaps in order to prevent errors from head misalignment. Data are transferred to and from the disk in sectors. Sector: Hundreds of sectors per track. Fixed or variable length with typically 512 bytes/sector. Adjacent sectors are separated by gaps. 4
Magnetic Disk Physical characteristics: Head Motion Fixed head (one/track) Movable head Platter Single platter Multiple platter Disk Portability Nonremovable disk Removable disk Head Mechanism Contact (floppy) Fixed gap Aerodynamic gap Sides Single sided Double sided 5
Magnetic Disk Magnetic disk: This type of memory is used for secondary storage. It is also called semirandomaccess memory, since the positioning of read/write transducers to the selected cylinder is random, and only the accessing of data within the selected track is sequential. 6
Magnetic Disk Disk performance parameters: The actual details depend on the computer system, the operating system, and the nature of the I/O channel and disk controller hardware. During disk operation: Disk is rotating at a constant speed to position the head at the desired track and at the beginning of the desired sector on that track. Track selection: Movable head system: moving the head, Fixed head system: electronically selecting one head. Seek time: For movable head system, is the time it takes to position the head at the track. Rotational delay/rotational latency: The time the disk tasks for the beginning sector to reach the head after the track has been found. 7
Magnetic Disk Disk performance parameters: Access time = Seek time + Rotational time Transfer time: the time required for the transfer of data (read/write) as the sector moves under the head. 8
RAID Redundant Array of Independent Disk: Rate in improvement in secondary storage performance < processors and main memory. High performance disk storage system Improvement of overall performance of computer system. Apply parallel processing concept to storage: Array of disks operates independently and in parallel, Higher performance can be obtained if required data reside on separate disk, Parallel I/O can be executed if blocks of data are distributed across multiple disks. Variety of ways can be done! Seven levels are proposed to be standard to make it possible to use in different application. 9
RAID Redundant Array of Independent Disk: RAID is a set of physical disk drives viewed by the OS as a single logical drive, Data are distributed across the physical drives of an array, Redundant disk capacity is used to store parity information, which guarantees data recoverability in case of disk failure [not supported in level 0]. 10
RAID-Level 0 RAID Level 0: Does not include redundancy to improve performance, User and system data are distributed across all of the disks in the array, Two I/O requests can be issued in parallel, if the requested blocks are on different disks, The data are striped across the available disks. Data are viewed as being stored on a logical disk which is divided into strips. For an n-disk array, upto n-logical strips can be requested in parallel. This greatly reduces the I/O transfer time. 11
RAID-Level 0 Logical disk Physical disk 0 Physical disk 1 Physical disk 2 Physical disk 3 Strip 0 Strip 1 Strip 2 Strip 3 Strip 4 Strip 5 Strip 6 Strip 7 Strip 8 Strip 9 Strip 10 Strip 11 Strip 12 Strip 13 Strip 14 Strip 15 Array Strip 0 Strip 4 Strip 8 Strip 12 Management Software Strip 1 Strip 5 Strip 9 Strip 13 Strip 2 Strip 6 Strip 10 Strip 14 Strip 3 Strip 7 Strip 11 Strip 15 12
RAID-Level 1 Redundant Array of Independent Disk: RAID Level 1: Data striping is also used. Each logical strip is mapped to 2 separate physical disks so that every disk in the array has a mirror disk that contains the same data. A read request can be serviced by either of the 2 disks: the one which has the minimum seek time + rotational latency, A write request requires that both corresponding strips be updated, but this can be done in parallel. Write performance depends on the disk with slower execution. Recovering from failure: When a drive fails, the data may still be accessed from the second drive. Drawback: Costly! It requires twice the disk space of the logical disk that it supports. Suitable for critical data storage. Strip 0 Strip 4 Strip 8 Strip 12 Strip 1 Strip 5 Strip 9 Strip 13 Strip 2 Strip 6 Strip 10 Strip 14 Strip 3 Strip 7 Strip 11 Strip 15 Strip 0 Strip 4 Strip 8 Strip 12 Strip 1 Strip 5 Strip 9 Strip 13 Strip 2 Strip 6 Strip 10 Strip 14 Mirror Strip 3 Strip 7 Strip 11 Strip 15 13
RAID-Level 2 Redundant Array of Independent Disk: RAID Level 2: Make use of parallel access technique, All member disks participate in the execution of every I/O request. The spindles of the individual drives are synchronized so that each disk head is in the same position on each disk at any given time. Data striping is also used with very small size of a single byte or word. A mechanism to correct error is implemented. Drawback: Still costly (less expensive than level 1) Strip 0 Strip 4 Strip 8 Strip 12 Strip 1 Strip 5 Strip 9 Strip 13 Strip 2 Strip 6 Strip 10 Strip 14 Strip 3 Strip 7 Strip 11 Strip 15 Strip 0 Strip 4 Strip 8 Strip 12 Strip 1 Strip 5 Strip 9 Strip 13 Strip 2 Strip 6 Strip 10 Strip 14 Error Correction 14
RAID-Level 3 Redundant Array of Independent Disk: RAID Level 3: Similar to level 2 but requires only a single redundant disk. Employ parallel access with small strip of data distribution. A simple parity bit is used to ensure data correctness. In the event of a drive failure, the parity drive is accessed and data is reconstructed from the remaining devices by use of the following approach Drawback: Still costly (less expensive than level 1) Strip 0 Strip 4 Strip 8 Strip 12 Strip 1 Strip 5 Strip 9 Strip 13 Strip 2 Strip 6 Strip 10 Strip 14 Strip 3 Strip 7 Strip 11 Strip 15 Error Correction 15
RAID-Level 3 Redundant Array of Independent Disk: RAID Level 3: Simple data reconstruction. Consider array of 5 drives: data are in {X0, X1, X2, X3} and parities are in X4. The parity for the ith bit is: X4(i) = X3(i) X2(i) X1(i) X3(i). Suppose that drive X1 has failed: X1(i) = X4(i) X3(i) X2(i) X0(i). Thus, the contents of each strip of data can be regenerated from the contents of the corresponding strips on the remaining disks in the array. This principle is true for RAID level 3 through 6. 16
RAID-Level 4 Redundant Array of Independent Disk: RAID Level 4: Make use of independent access technique. Each member disk operates independently. Separate I/O requests can be satisfied in parallel. Suitable for applications that require high I/O request rates. Data strip, with relatively large size, and parity technique are also used. Use concept of block-level parity. Block 0 Block 4 Block 8 Block 12 Block 1 Block 5 Block 9 Block 13 Block 2 Block 6 Block 10 Block 14 Block 3 Block 7 Block 11 Block 15 P(0-3) P(4-7) P(8-11) P(12-15) Error Correction 17
RAID-Level 5 Redundant Array of Independent Disk: RAID Level 5: Is organized in a similar fashion to RAID 4. Differs in the way that RAID 5 distributes the parity strips across all disks. By this way, the potential I/O bottleneck can be avoided. Block 0 Block 4 Block 8 Block 12 Block 1 Block 5 Block 9 P(12-15) Block 2 Block 6 P(8-11) Block 13 Block 3 P(4-7) Block 10 Block 14 P(0-3) Block 7 Block 11 Block 15 18
RAID-Level 6 Redundant Array of Independent Disk: RAID Level 6: Two different parity calculations are carried out and stored in separate blocks on different disks. The benefit of this technique is that it is possible to regenerate data even if two disks containing user data fail. Block 0 Block 4 Block 8 Block 12 Block 1 Block 5 Block 9 P(12-15) Block 2 Block 6 P(8-11) Q(12-15) Block 3 P(4-7) Q(8-11) Block 13 P(0-3) Q(4-7) Block 10 Block 14 Q(0-3) Block 7 Block 11 Block 15 19
RAID-Comparison Level Advantages Disadvantages Applications 0 I/O performance is greatly improved by spreading the I/O load across many channels and drives. No parity cal n overhead Very simple design Easy to implement 1 100% redundancy of data = no rebuild is necessary just a copy to the replacement disk Under certain cases, it can sustain multiple simultaneous drive failures Simplest subsystem design The failure of just one drive will result in all data in an array being lost. Highest disk overhead of all RAID types -- inefficient Video production and editing, Image editing, Any application requiring high bandwidth Accounting, Payroll, Financial, Any application requiring very high availability. 20
RAID-Comparison Level Advantages Disadvantages Applications 2 Extremely high data transfer rates possible The higher the data transfer rate required, the better the ratio of data disks to ECC disks. Relatively simple controller design. 3 Very high read data transfer rate Very high write data transfer rate Disk failure has an insignificant impact on throughput Low ratio of ECC disks to data disks high efficiency. Inefficient if very high ratio of ECC disks to data disks with smaller word sizes. Entry level cost very high. Requires very high transfer rate. Transaction rate equal to that of a single disk drive at best. Controller design is fairly complex. No comm implementation exist/not commercially viable Video production and live streaming, Image editing, Video editing, Any application requiring high throughput. 21
RAID-Comparison Level Advantages Disadvantages Applications 4 Very high read data transaction rate Low ratio of ECC disks to data disks means high efficiency 5 highest read data transaction rate Low ratio of ECC disks to data disks means high efficiency Good aggregate transfer rate Quite complex controller design. Worst transaction rate write Difficult and inefficient data rebuild in case of disk failure Most complex controller design. Difficult to rebuild in case of disk failure No comm implementation exist/not commercially viable File and application servers, Database server, Web, e-mail and news servers, Intranet servers, Most versatile RAID level 22
RAID-Comparison Level Advantages Disadvantages Applications 6 Provides for an extremely high data fault tolerance and can sustain multiple simultaneous drive failures More complex controller design. Controller overhead to compute parity addresses is extremely high Perfect solution for mission critical applications 23
Optical Memory Optical Memory: This storage media is a successor to the popular CD digital audio. Has a wide variety. CD-ROM: Very similar technology to audio CD. CD-ROM players are more rugged and have error correction devices to ensure that data are properly transferred to computer, Disk is formed from a resin (polycarbonate), Digitally recorded information is imprinted as a series of microscopic pits of the surface of the polycarbonate. The pits on the master disk are created by high-intensity, finely focused laser. The master is then used to make a die to stamp out copies onto polycarbonate. The pitted surface is then coated with a highly reflective material: Aluminium/Gold and protected with clear acrylic. Information is retrieved by a low-powered laser. The laser shines through the clear polycarbonate while a motor spins the disk past it. 24
Optical Memory Optical Memory: CD-ROM:. 25
Optical Memory Optical Memory: CD-ROM:. Protective Acrylic Label Polycarbonate Plastic Land Pit Aluminum Laser transmit/ receive 26
Optical Memory Optical Memory: CD-ROM:. 27
Optical Memory Optical Memory: CD-ROM:. Information does not organized on concentric track. A single spiral track is used. It begins near the center and spirals out to the outer edge. Equal size of sectors along the track. This makes the disk scans data with the same rate by rotating at variable speed. More slowly rotation for accessing the outer edge data. Data capacity is about 680MB. Block format: 2352 bytes Mode 0 = Blank data Mode 1 = 2048 bytes of data Mode 2 = 2336 bytes of data WO ECC 00 FF FF 00 Min Sec Sector Mode Data Layered ECC 12 bytes 4 bytes 2048 bytes 288 bytes SYNC ID Data L-ECC 10 bytes 28
Optical Memory Semi-reflective Optical Memory: DVD: Bits are packed more closely. The spacing between loop of spiral is reduced from 1.6 m to 0.834 m and also minimum distances between pits are reduced. It uses laser with shorter wavelength. Capacity is about 4.7 GB. It employs a second layer of pits and lands on top of the first layer. First layer is semi-reflective while second layer is reflective. By adjusting focus of laser, both layers can be accessed separately. To be more capable of storing data, DVD-ROM can be 2 sided. This makes the total capacity to be 17GB. 29
Magnetic Tape Magnetic Tape: The most common sequential-access memory device. A mylar tape coated with magnetic material on is used. Data is recorded as magnetic patterns. The tape moves past a read/write head to read or write data. There are two popular tape formats: The reel-to-reel tape used for storing large volumes of data (usually with largescale and mini-computer systems), The cassette tape used for small data volumes (usually with microcomputer systems) In both formats, data are recorded on tracks. A track on a magnetic tape runs along the length of the tape and occupied a width just sufficient to store a bit. On a nine-track tape, for example, the width of the tape is divided into nine tracks and each character of data is represented with 9 bits (one bit on each track). One or more of these bits is usually a parity bit which facilitates error detection and correction. 30
Magnetic Tape Magnetic Tape: Several characters grouped together form a record. The records are separated by an interrecord gap. A set of records form a file. The files are separated by an end-offile mark and a gap. 31
Magnetic Tape Magnetic Tape: Recording (writing): creating magnetic flux patterns on the device; The sensing of the flux pattern when the medium moves past the read/write head constitutes the reading of the data. In reel-to-reel tapes: data are recorded along the track. In a cassette tape: each bit is converted into an audio frequency and is recorded. It is noted that digital cassette recording techniques are also becoming popular. Magnetic tapes permit recording of vast amounts of data at very low cost. But the access time, being a function of the position of the data on the tape wrt. the read/write head position along the length of the tape, can be very long. 32
Magnetic Tape Magnetic Tape: The most common sequential-access memory device. A mylar tape coated with magnetic material on is used. Data is recorded as magnetic patterns. The tape moves past a read/write head to read or write data. There are two popular tape formats: The reel-to-reel tape used for storing large volumes of data (usually with largescale and mini-computer systems), The cassette tape used for small data volumes (usually with microcomputer systems) In both formats, data are recorded on tracks. A track on a magnetic tape runs along the length of the tape and occupied a width just sufficient to store a bit. On a nine-track tape, for example, the width of the tape is divided into nine tracks and each character of data is represented with 9 bits (one bit on each track). One or more of these bits is usually a parity bit which facilitates error detection and correction. 33