|Processor(s)||2x Intel Xeon Processor (Nocona core)3.6 GHz, FSB800, 1 MB L2 Cache|
|Platform||Asus NCL-DS (Socket 604)Intel E7520 Chipset, BIOS 1005|
|RAM||Corsair CM72DD512AR-400 (DDR2-400 ECC, reg.)2x 512 MB, CL3-3-3-10 Timings|
|System Hard Drive||Western Digital Caviar WD1200JB120 GB, 7200 RPM, 8 MB Cache, UltraATA/100|
|Mass Storage Controller(s)||Intel 82801EB UltraATA/100 Controller (ICH5)|
|Networking||Broadcom BCM5721 On-Board Gigabit Ethernet NIC|
|Graphics Card||On-Board GraphicsATI RageXL, 8 MB|
|Performance Measurement||c’t h2benchw 3.6|
|I/O Performance||IOMeter 2003.05.10Fileserver-BenchmarkWebserver-BenchmarkDatabase-BenchmarkWorkstation-Benchmark|
|System Software & Drivers|
|OS||Microsoft Windows Server 2003 Enterprise Edition,Service Pack 1|
|Platform Driver||Intel Chipset Installation Utility 188.8.131.525|
|Graphics Driver||Default Windows Graphics Driver|
The data compression offered by BrightStor is limited by CPU speed and the fact that the software is only single-threaded. In our test setup, the Dual Xeon had no impact on the time needed for backup or recovery. Our tests measured backup performance at 11-12 MB/s with compression enabled. Users wishing to archive entire drives will experience significant wait times, in our estimation about two hours and 15 minutes for 95 GB of data, which was the size of our test backup set. Without compression, data transfer rates are effectively doubled, leading one to wonder whether enabling software-based source coding is worthwhile.
As the chart illustrates, the original REV 35 had better backup performance than a Tandberg 220 LTO, which is a current tape backup device. Performance of the 35 GB edition almost matches the transfer rate churned out by the Tandberg 420 today, and the REV 70 easily beats the 35 GB REV. Enabling compression on the REV 35/70 slows transfer rates significantly, reducing backup performance in our tests by roughly 50%.
As mentioned earlier, one of the advantages of the REV backup solutions when compared to tape is that data access can be performed randomly instead of in a linear fashion. Users can also boot directly from a REV backup if need be. File seek times for REV 70 GB drives are much lower than those of tape, averaging approximately 12 ms for read and 13 ms for write. Access times for tape can be in excess of 75 seconds, depending on whether the data is located by convenient reference, and the physical proximity to the current tape/spool location.
Enabling compression has a significant impact on CPU load.