The Caen Knowledge Base is designed to provide information about
industry-specific terminology and basic implementation principles.
The below definitions were compiled and written by our RAID experts.
We will update and add definitions as new technology evolves.
Cables
and Cabling
Cables can be the single most important piece of a successful
system. Throughput and data transfer reliability are dependent
upon the interconnectivity between major components. As a result,
proper cabling can make or break any system. One of the major
concerns with cabling is impedance (inability to transmit a clear
signal). Signals are limited depending on strength, type and cable
material.
The following list displays maximum recommended cable lengths
(devices and cabling, both internal and external, must be factored
into calculations; these calculations are based on a system with
8 devices):
 Single
Ended SCSI: 6 Meters / 19.7 feet
Single
Ended Fast SCSI: 3 Meters / 9.8 feet Ultra Wide
Single
Ended SCSI: 1.5 Meters / 4.9 feet
Ultra
High Voltage Differential (HVD) SCSI: 25 Meters / 82 feet
Low
Voltage Differential (LVD) SCSI: 12 Meters / 39.4 feet
Disk Mirror (RAID Level 1)
A disk mirror is a complete copy of data that resides on one
physical disk to another physical disk.
Advantages:
Better
read performance than Parity RAID.
Moderate
write performance increase.
Higher
fault tolerance (a higher percentage of disks in a mirrored redundancy
group may fail simultaneously as compared to a parity RAID redundancy
group).
Disadvantages:
This
option is very expensive because it requires nearly double the
amount of data storage hardware to facilitate its implementation.
Disk Striping
Disk striping is a technique used for spreading data over multiple
disk drives. Disk striping can speed up operations that retrieve
data from disk storage. The RAID controller breaks a body of data
into units and spreads these units across the available disks
in the RAID set. Disk striping stores each data unit in only one
place and does not offer protection from disk failure. It is not
a "true" RAID because it is not fault tolerant.
Fault Tolerance
Fault tolerance is a system's ability to remain operational
in the event of a component, device or environmental failure.
For example, if the hard disk fails in a single hard disk system,
the system goes off-line (you have a serious problem). However,
if one hard disk drive fails in a RAID system, the data that is
stored on that individual disk drive is redundant and recovery
from the fault is possible and the system remains on-line.
Fibre Channel
Fibre Channel technology is a communications protocol designed
to meet requirements related to the increasing demand for high
performance information transfer. Fibre Channel attempts to combine
the benefits of both channel and network technologies. The goals
of Fibre Channel include:
Support
for multiple data rates, media types, interfaces and connectors
High
bandwidth (100MB/s and beyond)
Flexible
topologies
Reliable
connectivity over several kilometers
Fibre Channel is making the biggest impact in the storage arena,
especially when employed using SCSI as its upper layer protocol.
find out more about
Fibre | buy
Fibre interface products
FRU
A FRU is a Field Replaceable Unit. Think of it as a spare part
that can be swapped out with a failed component. Typical FRUs
are hot-swap hard disk drive canisters, power supplies, and fan
assemblies. Caen offers hot-swap controllers and various types
of component mounting accessories to facilitate the expedient
replacement of failed devices.
Host
A host is a parent or base system that is accessing the RAID
array for the purpose of data storage.
HVD
HVD stands for High Voltage Differential.
LUN
LUN stands for Logical Unit Number. A LUN is a method of expanding
the number of devices that can be placed on one bus domain. Each
LUN can address up to eight devices at each SCSI ID.
LVD
LVD is an acronym for Low Voltage Differential and is a subset
of Ultra2 SCSI technology. LVD uses differential signaling technology
which has lower voltage swings and is less susceptible to noise
than standard Ultra SCSI technology.
find out more about LVD
Parity RAID (RAID Levels 3, 4, 5, and 6)
Unlike full mirroring (100% of the data is duplicated), parity
RAID reduces the percentage of mirrored data to 10-33%. In parity
RAID, not all data is copied verbatim. You can think of it as
a type of compression or abbreviation of the actual data that
is made redundant.
Advantages:
Less
overhead in the amount of data being used to provide redundancy.
More
cost effective than mirroring.
If
a second disk fails at the same time or during the reconstruction
phase of the RAID set you can still have the ability to recover
successfully.
Disadvantages:
Varying
degrees of performance degradation depending upon the manner in
which data and redundant data are mapped to the disks comprising
the RAID array. In RAID levels 3-5, if more than one disk fails
at any given time or during reconstruction of the RAID set after
a failure of another disk, you have a serious problem. Why?
RAID
RAID stands for Redundant Array of Independent Disks (also called
Redundant Array of Inexpensive Disks). RAID is a method of enabling
several physical hard disk drives to act as a single orchestrated
storage area. Using RAID can achieve fault tolerance in the event
of a disk drive failure. It also can provide higher throughput
levels than a single hard drive or group of independently configured
hard drives. RAID also allows for uninterrupted operation and
automatic reconstruction of the data contained on the failed disk
drive(s) to ensure that productivity continues in the event of
hardware or environmental malfunction and/or failure.
buy PATA
RAID solutions
RAID 0
RAID 0 is a Striped Disk Array with No Fault Tolerance. Disk
Striping only (which is used by both Parity RAID and often by
Mirroring RAID) does not contribute to fault tolerance or the
ability to recover after a catastrophic failure. It simply implies
that data is spread across several different physical disks.
RAID 0+1
RAID 0+1 is a Striped Disk Array with Fault Tolerance. It is
a combination of the disk mirroring features of RAID 1 with the
disk striping features of RAID 0 and allows for increased speed.
RAID 1
Mirroring/Duplexing. Provides disk Mirroring.
RAID 2
Hamming Code ECC. Requires the use of non-standard disk drives-not
viable for commercial use.
RAID 3
RAID 3 is a parallel transfer to multiple disks with a separate
shared parity disk. Same as RAID 0, but also reserves one dedicated
disk for error correction data. It provides good performance and
some level of fault tolerance. Different from other Parity RAID
levels (4, 5, & 6) because disks in a RAID level 3 RAID array
operate in parallel or unison (as opposed to independently). This
difference makes RAID level 3 best suited for high-bandwidth applications.
RAID 4
Independent Data Disks with a Separate Shared Parity Disk.
RAID 5
RAID 5 are independent data disks with distributed parity blocks
on each disk. Provides data striping at the byte level and also
stripe error correction information. This results in excellent
performance and good fault tolerance.
RAID 6
RAID 6 are independent data disks with two independent distributed
parity schemes on each disk.
Serial ATA
Ultra ATA is the primary internal storage interconnect for the
desktop, connecting the host system to peripherals such as hard
drives, optical drives, and removable magnetic media devices.
Serial ATA is the next -generation internal storage interconnect
designed to replace Ultra ATA technology. Serial ATA is the proactive
evolution of the ATA interface from a parallel bus to a serial
bus architecture. This architecture overcomes the electrical constraints
that are increasing the difficulty of continued speed enhancements
for the classic parallel ATA bus. Serial ATA will be introduced
at 150Mbytes/sec, with a roadmap already planned to 600Mbytes/sec,
supporting up to 10 years of storage evolution based on historical
trends. Though Serial ATA will not be able to directly interface
with legacy Ultra ATA hardware, it is fully compliant with the
ATA protocol and thus is software compatible.
See the SATA
website (external link)
SCA
SCA stands for Single Connector Attachment. Used with direct
connect disk drives to eliminate additional cabling, SCA is extremely
helpful when using LVD connections and is a way to reduce impedance
caused by extra connecting cables and hardware.
SCSI
SCSI stands for Small Computer System Interface. Pronounced
"Skuzzy," SCSI is a parallel interface standard that is used by
PC's, Apple Macintosh computers, and many UNIX systems for use
in connecting peripheral devices. SCSI provides faster data transmission
rates than slower standard parallel ports.
find out more about
SCSI | buy
SCSI interface products
SCSI-1
8-Bit
Bus size
Uses
a standard 25-pin connector
Supports
data throughput rates of up to 4 MBps
SCSI-2
8-Bit
Bus size
Uses
a 50-pin connector
Supports
data throughput rates of up to 4 MBps
SCSI-3 (Ultra Wide SCSI)
16-Bit
Bus size
Uses
several different connectors
Supports
data throughput rates of up to 40 MBps
SCSI-Fast
8-Bit
Bus size
Uses
several different connectors
Supports
data throughput rates of up to 10 MBps because it doubles the
clock rate
SCSI-Fast/Wide
16-Bit
Bus size
Uses
several different connectors
Supports
data throughput rates of up to 20 MBps
SCSI-Ultra
8-Bit
Bus size
Uses
several different connectors
Supports
data throughput rates of up to 20 MBps
SCSI-Ultra2 Wide
16-Bit
Bus size
Uses
several different connectors
Supports
data throughput rates of up to 80 MBps
SCSI-Ultra320 Wide
16-Bit
Bus size
Uses
several different connectors
Supports
data throughput rates of up to 160 MBps
SCSI-Wide
16-Bit
Bus size
Uses
a 68pin cable
Supports
data throughput rates of up to 4 MBps
|
