Compaq ProLiant 6000 Compaq DLT Tape Array II: High-Performance Backup of Ente - Page 21

Ersus, Indows

Get Compaq ProLiant 6000 PDF manuals and user guides View all Compaq ProLiant 6000 manuals
Add to My Manuals
Save this manual to your list of manuals

Page 21 highlights

ECG075.0997 (cont.) DLT Tape Array II ... TEST 7 RAIT-0 VERSUS RAIT-5  WINDOWS NT Windows NT 4.0 Cheyenne ARCserve 6.0 for Windows NT Image  single job 2 GB 2:1, 4:1 RAIT-0 and RAIT-5 ProLiant 5000, two Pentium Pro 200 processors, 256K cache, 256 MB RAM 80-GB array (two SMART-2/P Array Controllers with twenty 4-GB drives) 35/70 GB DLT drive(s) Wide-Ultra SCSI-3 cards (All Compaq off-the-shelf products) To determine the impact of RAIT-5 as it relates to processor performance and throughput, to discover how much CPU power is required to perform the RAIT-5 calculations, and to find out whether this overhead will seriously affect backup times. Tests show that the performance of RAIT-5 in a typical configuration (two SCSI cards connected to four drives) is only 1 to 7 percent slower (depending on data set compressibility) than the RAIT-0 backups. This suggests that RAIT-5 should be considered the optimum choice in a decision between speed and safety. RAIT-5 not only writes to a set of tapes as an array, but uses a parity stripe to provide checksum information, which permits file restoration even when there has been some data loss on a single tape in the array. On the other hand, RAIT-0 connects the drives together logically and lets them be written to as a single drive. Because RAIT-0 does not perform the striping calculations, it is faster than RAIT-5. There is a significant increase in CPU use for RAIT-5 over RAIT-0 (7 percent versus 41 percent), which could be a problem with concurrent backups or backups made during heavy server activity. The impact of RAIT-5 on server resources can be reduced by adding processors to the system.

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
7
%
(cont.)
DLT Tape Array II
ECG075.0997
T
EST
7
RAIT-0 V
ERSUS
RAIT-5
W
INDOWS
NT
Windows NT 4.0
Cheyenne ARCserve 6.0 for Windows NT
Image
single job
2 GB
2:1, 4:1
RAIT-0 and RAIT-5
ProLiant 5000, two Pentium Pro 200 processors, 256K
cache, 256 MB RAM
80-GB array (two SMART-2/P Array Controllers
with twenty 4-GB drives)
35/70 GB DLT drive(s)
Wide-Ultra SCSI-3 cards
(All Compaq off-the-shelf products)
To determine the impact of RAIT-5 as it relates to processor performance and throughput, to
discover how much CPU power is required to perform the RAIT-5 calculations, and to find out
whether this overhead will seriously affect backup times.
Tests show that the performance of RAIT-5 in a typical configuration (two SCSI cards
connected to four drives) is only 1 to 7 percent slower (depending on data set compressibility)
than the RAIT-0 backups. This suggests that RAIT-5 should be considered the optimum choice
in a decision between speed and safety. RAIT-5 not only writes to a set of tapes as an array, but
uses a parity stripe to provide checksum information, which permits file restoration even when
there has been some data loss on a single tape in the array.
On the other hand, RAIT-0 connects the drives together logically and lets them be written to as
a single drive. Because RAIT-0 does not perform the striping calculations, it is faster than
RAIT-5. There is a significant increase in CPU use for RAIT-5 over RAIT-0 (7 percent versus
41 percent), which could be a problem with concurrent backups or backups made during heavy
server activity. The impact of RAIT-5 on server resources can be reduced by adding processors
to the system.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.