Chapter 9 Introduction to the Server
385
fault tolerance
as applied to RAID
systems, it is the
ability to recover
from a hard disk or
hard disk controller
failure without the
loss of stored data. As
applied to a network
infrastructure, it is
the ability to continue
operation during a
system hardware or
software error.
disk striping
a storage technique
that divides the data
into sections and
writes the data across
several hard disk
drives at the same
time.
error correction
a RAID technique
that uses traditional
error-checking code
(ECC) or parity.
Another RAID technique is error correction, which can use traditional
error-checking code (ECC) or parity. ECC is obsolete as a method for RAID
because most hard disk drive systems use ECC as a standard way to protect data.
Incorporating ECC into a RAID system would be redundant.
Parity is a technique that allows data to be recovered if one of the hard disk
drives fails in a multiple disk drive system. A minimum of three hard disk drives
must be used for this technique. If three hard disk drives are used, one hard
disk drive is used to store parity and the other two are used to store data. Data
is spread evenly between the two data storage drives, and the parity code for the
sum of the two data storage drives is stored on the parity drive.
Parity is also a binary code that represents the total data pattern shared
between the data storage drives. If any data storage drive fails, the system can use
the parity bit to rebuild the missing data.
RAID systems are referred to as RAID levels. There are many RAID levels,
each incorporating a different technique to increase speed or to provide data
redundancy, or both.
Things to consider when selecting a RAID level are cost, reliability, and
speed. There is no single, best RAID level. The appropriate level depends on the
value of the data being protected and the amount of budget available. If a bank
were protecting fi nancial data, cost would most likely be of little concern. A
RAID 0/5 with a tape backup would be quite appropriate. If you were protecting
a personal computer, you would not likely need a RAID system.
Repairing a failed RAID system means the drive must be physically replaced
and then the data reconstructed. In a Microsoft Windows operating system, the
Disk Management utility is used to reconstruct data. If using proprietary SCSI
drives, data is reconstructed by the data array manufacturer software package.
Figure 9-18.
Memory
assignments.