User guide

12 Us er’s Guide
 A memory address space of 16 MB
(80286 microprocessor) to 4 GB
(Intel386 or higher microprocessor)
 Multitasking
 Virtual memory, a method for
increasing addressable memory by
using the hard-disk drive
The Windows NT 32-bit operating system
runs in protected mode. MS-DOS cannot
run in protected mode; however, some
programs that you can start from
MS-DOS—such as Windows—are able to
put the computer into protected mode.
/-
Abbreviation for Personal System/2.

Abbreviation for power-supply paralleling
board.
8
Abbreviation for polyvinyl chloride.
1
Abbreviation for quarter-inch cartridge.
'
Acronym for redundant array of indepen-
dent disks. This phrase was introduced by
David Patterson, Garth Gibson, and Randy
Katz at the University of California at Ber-
keley in 1987. The goal of RAID is to use
multiple small, inexpensive disk drives to
provide high storage capacity and perfor-
mance while maintaining or improving the
reliability of the disk subsystem.
Patterson, Gibson, and Katz described
five different methods, which are known
as RAID levels 1 through 5. Each level
uses one or more extra drives to provide
a means of recovering data lost when a
disk fails, so that the effective failure rate
of the whole disk subsystem becomes
very low.
'.
RAID 0 is commonly called striping. This
was not originally defined as a RAID level
but has since come into popular use. In
this array configuration, data is written se-
quentially across the available disks and
no redundancy is provided. RAID 0 config-
urations provide very high performance
but relatively low reliability. RAID 0 is the
best choice when controller cards are du-
plexed. See also striping.
'9
RAID 1 is commonly called mirroring.
RAID 1 also uses striping, so RAID 1 may
be regarded as the mirroring of RAID 0
configurations. RAID 1 is the best choice
in high-availability applications that require
high performance or relatively low data ca-
pacity. See also mirroring, RAID 10,
striping.
':
RAID 4 is commonly called guarding. It
uses data striping, like RAID 0, but adds a
single, dedicated parity drive. The parity
data stored on this drive can be used to
recover data lost from a single failed drive.
RAID 4 configurations write data slowly
because parity data has to be generated
and written to the parity drive, and the
generation of the parity data frequently re-
quires reading data from multiple physical
drives. See also guarding and striping.
';
RAID 5, like RAID 4, is commonly called
guarding. RAID 5 is identical to RAID 4,
except that the parity data is distributed
evenly across all physical drives instead
of a parity drive. In configurations using a
large number of physical drives in which
a large number of simultaneous small
write operations are being performed,
RAID 5 offers potentially higher perfor-
mance than RAID 4. RAID 4 and RAID 5
configurations are appropriate in high-
availability applications where
performance is less critical or where high
data capacity is required. See also
guarding.
'9.
RAID 10 is a mirroring technique in which
data is duplicated across two identical
RAID 0 array or hard-disk drives. All data
on a physical drive in one array is duplicat-
ed, or mirrored, on a drive in the second
array. Mirroring offers complete redun-
dancy of data for greater data security.
See also mirroring, RAID 1, and striping.