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