RAB Disk Classification
Program
These are the RAID Advisory Board (RAB)
standards (basic and premium grades), for comparing the
integrity of RAID installations: FRDS, FTDS, and DTDS.
Failure Resistant Disk Systems (FRDSes), installations are designed to prevent data loss or system failure caused by the failure of any single RAID drive.
FRDS+, the premium grade, prevents data loss caused by high operating temperatures, external power failure or even disk cache failure. It even supports the automatic "hot" swap of drives while the system is poweredon.
Failure Tolerant Disk Systems (FTDSes), installations insure continuous data availablity despite the failure of any single system component; as well as, the failure of the RAID's server.
FTDS+, the premium grade, provides data-addess protection against system's I/O bus and loss of external power. It also supports the automatic "hot" swap of drives while the system is poweredon.
Disaster Tolerant Disk Systems (DTDSes), installations are divided into two or more sites or "zones, " either in the same building or in different facilities, but with separate power sources. If the RAID in one zone fails--perhaps due to its failure or a failure in its server system, or even the facility's cooling system--the other RAID is still online. It provides protect against data loss due to power outages.
DTDS+, the premium grade, separates DTDS zones by at least one kilometer while maintaining data consistency. DTDS+ offers protection against massive equipment and environmental failures including blackouts, fires, floods, and vandalism and theft.
RAID
Redundant Array of Inexpensive or Independent (if not proprietary) Drives. Two or more drives work in tandem to provide redundant data storage. If one of the drives fails, stored data can still be reconstructed from the other drive or drives that still operate&emdash;as if the failure had not occurred. When capacity, speed and reliability are of utmost consideration, as they are for large local area networks (LANS) and wide area networks (WANS), a managed "array" of hard drives is the preferred security consideration. With many drives sharing data-storage and retrieval operations, the mechanical stress on individual drives is reduced&emdash;the mean time before failure (MTBF) for each drive is extended beyond the 150,000 hour average for quality drives. There are six options or "levels" of implementation, RAID-0 to RAID-5.
RAID-0. Here, software manages the mirroring process among daisy-chained drives, usually they are SCSI-interfaced. Communications "throughput" between these devices is as fast as the SCSI-interface and control implementation, or any alternative proprietary connection. However, there is no fault tolerance&emdash;if one device fails, the others are no longer available.
RAID-1 is a mirroring process at the sector level or lower, with a single controllerboard managing multiple hard drives in most implementations. Data bytes are stripped and spread among the drives, but without an advanced error-checking protocol. RAID-2 is a duplexing setup. To each data byte, parity bits are added, then data is interleaved--placed by mathematical formula--among multiple data drives. Two or more "redundant" parity drives (or disks in each drive) store complementary parity information. (This "Hemming" error-detection system is similar to its implementation for serial-interfaced devices.) Parity codes provide error detection and correction. Typically, to support 10 "working" data drives (or disks), 4 others are dedicated to parity.
RAID-2 is fast for large file transfers, but less efficient for small file transfers, because all drives in the array are accessed for every request. RAID-3, -4, and -5 are also duplex arrangements; however, drives are tied together into a single subsystem of "disk channels" for a more rigidly controlled and efficient process. Using special techniques to spread data among the drives, and to detect and correct errors as they occur, these RAIDs provides higher performance than many single drives&endash;with far greater reliability. To the computer, a RAID appears as a single controllerboard or (SCSI or IDE) interfaceboard. Drive thoughput, for any manufacturer's implementation of a RAID is also determined by the application, block size of the data transfers, and operating system.
RAID-3 implementations use an array of drives: two or more data drives and one additional "redundant" parity drive. Data is transferred in parallel to data and parity drives. It is spread ("byte striped") by formula among the data drives. Parity bits, generated by a mathematical formula called an exclusive OR (XOR), are included among them. The parity drive stores a complementary copy of the parity information When a data drive fails, the remaining data drives and the parity drive together can recreate the missing information.
An array's writes to disk are slowed down by the error-correction calculations; reads are much faster&endash;unless an error is detected. Access time is determined by the number of drives in the array. Because it takes less time to only send part of a data byte to each drive, the more drives in an array, the faster the throughput. Ideally an array's performance approaches the limits of its interface's capability which is far greater than the real world performance of a single drive. Typically, a single drive's access is reduced by mechanical factors; for example, an Ultra-wide SCSI-3 interface is capable of 40MByte-per-second (40MBps) transfers, but single drives rarely approach 20MBps using this interface.
RAID-4 has multiple data drives supported by a single redundant parity checking drive. Data sectors are interleaved between disks (instead of bytes as in RAID-3). While only a single write can take place at any time, because the parity drive must be updated, multiple concurrent reads from two or more drives are supported. Therefore, many more read requests can be handled per second than writes.
RAID-5 is similar to RAID-4, but its error-correction codes are stored on each drive in the array. Because parity information is on each drive, this array can support concurrent read and writes to multiple drives. The key is the ability of the controller to optimize the process; for such implementations, the write performance may be slower than RAID-1 or RAID-3. Also, refer to the RAB Disk Clasification Program.
RAIL
Redundant Array of Inexpensive Libraries, similar to a RAID device, but composed of DVD-RAM drives instead of (more than one) hard drives. RAILs have mechanical magazines with a capacity for two or more DVDs; which typically, can be changed while the unit is poweredon.
RAM
Random Access Memory chips. RAM is the memory in your computer where applications and the operating systems run. Your computer is able to write and read data very quickly to and from RAM as opposed to a hard drive. When you launch an application it loads as much of itself as possible into RAM. This allows the applications to execute as quickly as possible.
RAM is volatile, data in RAM must be saved to data storage media or it will be lost when the RAM chips lose power. This takes the information out of the temporary high-speed memory (RAM) and writes it to the slower but permanent memory (hard drive). There are several types of RAM chips which are also referenced: DRAM, Pseudo Static RAM, Static RAM and Flash RAM. Note: PC What's The Problem includes functional troubleshooting suggestions.
RAM Disk A RAM Disk is a term used to represent a logical structure of allocated RAM (chips or SIMMs) capacity which simulates a data-stroage drive, albeit one that is performing at much higher speeds than a data-stroage drive is capable of acheiving. A RAM Disk is volatile, the chips require power to maintain data. RAM Disks are used for applications where otherwise there would be constant disk accessing such as databases and for Internet browser caches. Also refer to Virtual memory and Flash
RDRAM
Rambus DRAM is a DRAM with a proprietary clocked bytewide Rambus Channel. It operates at 250 MHz and uses both clock edges for synchronous data transfers. RDRAM has a 500 MByte-per-second data access rate; but, random accesses are much slower. By 1999, RDRAMs should support 800MHz RIMM Inline Memorie Modules to provide a maximum bandwidth of 1.5 Gbytes per second. RIMM sockets are similar in size to DIMM sockets but they are not compatible. These modules consume roughly 2 watts of power.
Read/Write Heads (Hds)
This electromagnetic device both reads information stored on a disk-platter and writes to store information on the disk. Inside a drive, each disk usually has two Read/write heads, one for each side ("platter"), top and bottom. A drive with eight disks will have sixteen heads, or fifteen heads if that is all that the manufacturer could fit into a drive with the specified height. All heads are moved, at the same time, to the same position over their assigned platters. Each head is fixed to the extended end of a metal "acutator-arm." There are two types of mechanisms to move the actuator-arms: Stepper Motor and a Voice Coil.
Rectifier diodes
These are referenced with all others as: Diodes.
Red Book Audio
Audio Engineering Society's file format specification for audio CDs supporting a maximum bit depth of 16 and a sample rate of 44.1K.
Reduced Write Current (RWC), Write Precomp (WP)
These references apply to older hard drive designs. Write Precomp(ensation) (WP) is the timing of the electrical current used by a read/write head to record data. It varies by each track's circumference so that the same number of bits can be packed into the narrower sectors on tracks near the hub as in the wider sectors outward to the rim of the disk. Starting with the specified sector and inwards towards the hub the signal timing increases to compensate for the narrow sectors.
Write Current or "Reduced Write Current" is the amplitude of the electrical signal used by a read/write head to record data. The signal is reduced for writing to inner track sectors starting with the specified sector and inwards towards the hub. In these sectors, data bits are more densely packed and using a stronger signal to record a bit would affect any adjacently stored bits. The numbers indicating both the start of WP and the start of reduced WC are not used when they equal the number of cylinders plus 1.
Remote File Transfer
Permits file transfer, as well as access to the resources of the remote system, including its application programs and operating system. Other file transfer references: File Transfer - Basic Capture, File Transfer - Basic Send, Full-Feature File Transfer, and Unattended-Remote File Transfer.
Resistor
An electrical component that decreases the amount of current flowing through it. Composed of carbon compounds that resist the flow of electrons, they are used to protect sensitive circuits. Resistors with very high capacities have their capacities labeled on their sides. Computers and related equipment use Lower-capacity resistors.
These resistors use colored bands to indicate their properties. The first band in from one side indicates its electrical tolerance. The two or three bands in from the opposite side indicate the values of the first, second, and (for high-precision resistors) third significant digit. The band closest to the "tolerance" band, known as the multiplier band, indicates the number of zeros to be placed after the significant digits. Example: to represent a value of 331, the bands would be orange, orange, black. The last band is for precision.
Wire-wound (WW) resistors are stable over time and can dissipate a lot of power, but are unsuitable for RF (radio frequency) use and temperature stability is less than that of metal-film resistors.
Metal-film (MF) resistors have good thermal stability and shelf life, and low noise levels and in general are a better choice than carbon-film (CF) resistors. The best "precision" MF resistors should always be replaced by similar components.
Carbon-Composition (CC) resistors are rarely used because their random movement of charge causes high noise levels, soldering variances can change their values, and they are sensitive to hot and humid environments.
Variable resistors, such as a potentiometer, reostat, trimmer, etc., may use a sliding contact, or brush, touching the resistance-media to increase or decrease resistance.
There are many types of resistance media, as previously noted. Computers and related equipment often use resistors made from carbon compounds and conductive plastic. Large wattages require a high-capacity resistor known as a rheostat; these power regulators use wire with special electrical-resistant properties.
Light-dependent resistor. Its symbol in a schematic drawing uses two slanting arrows to represent the direction of light from its source towards a photoresistor:
Potentiometer or "pot" is a variable resistor that uses a sliding contact (or wiper arm) to increase or decrease resistance. Typically, the sliding contact is moved by either a knob or a round slotted head which can be turned with either your fingers or a screwdriver. For example, a light-dimmer uses a potentiometer to increase or decrease the amount of illumination that a light bulb produces by decreasing or increasing the resistance to the electricity flowing to the light bulb. Cerment pots have similar reliability to carbon pots and can serve as substitutes for the original. (For even better reliability, use conductive plastic resistors.)
RHD
Removable hard drive (rhd) which includes its own logic controls, drive motor and read/write head mechanisms inside the removable unit.
RHS
Right Hand Side. Refers to the arithmetic or transfer part of a processor's internal cluster.
RISC
Reduced Instruction Set Computer's processor.
RJ-11 and RJ-45
The Universal Service Ordering Code (USOC) specifies over 60 Register Jack (RJ) cable and connector configurations. For voice and data transmission, not connected to the public telephone network, generic RJ-11 and RJ-45 should work. For RJ-11 Cables and connections, most indoor telephone systems use unshielded 4-wire wall/station cable. They can support only software handshaking. A dedicated control wire is required for hardware handshaking.
RJ-12 Cables and connections used for interior communications and inside PBX systems, use shielded, stranded (rather than solid), multi-paired (3 pairs) cable. Both software and hardware handshaking are supportable. Distances longer than 3,000 feet may require a repeater.
RJ-45 Cables and connectors, supporting electronic communications require cables that are shielded against both electromagnetic interference (EMI) and radio frequency interference (RFI). Choose cable with the best flexibility and strength available.
Note: PC What's The Problem? includes identification, test, repair and replacement suggestions.
RLL
All ESDI drives use RLL data-encoding and have 34 or more sectors per track. Some ESDI drives use "RLL/Non," non return to zero; and have up to 60 sectors per track. ESDI drives require controllers with more precision than ST506 drives. If you match an ST506 controller with a RLL drive, you may be limited to 26 sectors per track. RLL drives should be "RLL-certified."
To store more data, a RLL drive requires more accurate read/write head mechanisms than MFM drives. It needs such heads to write data to and read data from tracks of higher density. (The tracks are thinner and there are more of them.) On the disk's tracks, the heads create bit cells (magnetic domains) that are one-third smaller than used in MFM encoding. RLL controllers use a special code that can read 16-bit patterns of information, rather than the 8-bit patterns of MFM.
There are two RLL coding techniques which are differentiated by how many consecutive zero bits can be stored before automatically inserting a one bit. (Bits are stored as either 1 or 0.) RLL (2,7) allows from two to seven consecutive zero bits. RLL (3,9) allows from three to nine consecutive zero bits. RLL (2,7) provides a 50% increase in storage capacity over standard MFM; RLL (3,9) provides 100%.
To address the aspects of signal degradation caused by nonlinear effects and to increase data density, some drives use RLL 1,7 (d=1, k=7), with a minimum of one zero between transitions to address partial erasure and no more than 7 zeros in a row to assist in clock-sync recovery.
Router
A network communications device designed to transmit signals via the most efficient route possible.
rpm
Rotations per minute. Drives of older Winchester designs spin their disks at 3,600 rpm. Mostly in response to multimedia AV (audio/video) requirements, high-speed interfaces, data-storage caches&emdash;manufacturers have increased drives to 4,500, 5,400 7,200 and 10,000 rpms as well as the efficiency of the read/write heads.
rtz
Return to Track Zero function recalibrated the starting position of read/write heads by moving them to the first track in from the rim, track 0 after compleating a read or write operation.
References are samples only. Each one is presented in greater detail in the
Technical Research Assistant for 2001
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Did you find any Inaccuracies?
Your EMAIL comments are invited
on CD in Adobe Acrobat PDF format
Book
Visit our
Enter the library through its Opening page
All rights reserved.
to top