Contact Information
500D Huntmar Park Drive
Herndon, VA 20170
Phone: (703) 471-8108
Fax: (703) 471-8195
techsupport@risingedge.com
|
RAID Benefits
RAID provides three key benefits - fault tolerance, data availability, and high
performance. These three benefits combined, result in unparalleled storage subsystem
reliability and performance.
Fault Tolerance
The ability of a mass storage subsystem to sustain a component failure without data loss
is known as fault tolerance. With non-fault tolerant storage devices, a component failure
typically results in the loss of data. This is the case in the all too familiar disk
crash. The fault tolerance provided by RAID however, will insulate users from this form
of data loss. Since the data on a RAID system is stored redundantly, the loss of a disk
drive within the array does not result in the loss of data. Rather, the data stored on
the remaining functional drives can be used to reconstruct the data from the failed drive.
Data Availability
The ability of a mass storage subsystem to sustain a component failure without
loss of functionality is known as data availability. Fault tolerance leads directly to
data availability. Without storage subsystem fault tolerance, a failed storage component
results in storage subsystem failure which in turn results in a data-down condition, or a
lack of data availability. This lack of data availability extends for the duration of
failure isolation, repair and data reconstruction. The fault tolerance provided by RAID
removes the data-down condition by continuing to function while the failed component is
replaced. The redundant data stored on the remaining drives in the array are used to
reconstruct the data stored on the failed drive during this repair period.
Data backups are not a substitute for fault tolerance and data availability.
Since a backup is typically not up-to-date with current working data, all data between the
time of the last backup and component failure is lost. Also, the system down time
associated with isolating the failure, repairing the system, and restoring the data can
easily run into hours or days and thousands or hundreds of thousands of dollars in lost
time.
The fault tolerance and data availability provided by RAID combine to provide a
highly reliable mass storage subsystem. Reliability is measured as Mean Time Between Data
Loss (MTBDL) which is the mean time to a failure that will result in data loss or data
unavailability. The high reliability of today’s disk drive significantly impacts MTBDL.
For example, a typical Mean Time To Failure (MTTF) for a commercially available disk drive
is between 30,000 and 100,000 hours. Given a RAID 5 system using 100,000 MTTF disk drives,
MTBDL is equal to 2,500,000 hours or 285 years!
High Performance
High performance is the ability of a mass storage subsystem to perform I/O and data
transfer requests at high rates. RAID systems offer significant performance advantages
over other storage subsystems. RAID systems typically fall into one of two categories -
independent access arrays or parallel access arrays. Independent access arrays, as the
name implies, access each member disk independently. This allows the RAID to handle
multiple IO requests for data that is physically stored on different member disks within
the array at the same time. Parallel access array access all member drives within the
array simultaneously. That is, a block of data is stored on all data disks within the
array by splitting the original data received from the host into smaller blocks that are
written to each data disk within the array. Conversely, data read from the RAID consists
of smaller blocks of data being read from each member data disk. This parallel disk
access results in an aggregate data that approaches the sum of the data rate for all
member data disk drives within the array.
|
|
