Seagate 7200.11 (Raid 0 Results)
Author: Dominick V. Strippoli
We recently reviewed the latest drive from Seagate's arsenal: Known as the 7200.11 Barracuda. This drive proved to be an excellent competitor for the well known Western Digital Raptor 150. General performance of the drive was very comparable to the Raptor and ended up about 7% to 15% slower in every day tasks. On the other hand, when working with video/photo editing and larger file operations the Seagate 7200.11 had a very sizable 25% to 30% advantage. As highly requested by our readers, we have put together a very simple review article of the same 7200.11 drive used in a Raid 0 array. We will use the same testing methodology as our previous review and simply add one additional Barracuda to create a Raid 0 array. Once again, we will be using the Western Digital Raptor 150 as a comparison drive. In order to obtain the maximum possible sustained transfer rates on the drives, and the quickest possible access time we have limited our Raid 0 volume to a 50GB size and 128k stripe. This in turn will only use the first 25GB's (and fastest portions) of each drive. The remaining 440GB's or so, for the purpose of taking advantage of the Matrix Storage Controller we have chosen to create a mirrored Raid 1 setup. Since we have already reviewed the single Seagate 7200.11 in a full scale in-depth review, we are simply going to provide this introduction, a small conclusion, and solely hard results from our testing.
Synthetic Testing Results
Raid 0 HDTach (50 GB total, 2 X 25GB)
Raid 1 HDTach (440GB Mirrored)
WEB SERVER TEST (100% Random Read)
DATABASE SERVER TEST (100% Random Write)
Real World Testing
In conclusion, we see that the raid array can indeed prove to be beneficial for task specific users. Let us compare the Raid 0 Barracuda setup to the single Barracuda setup. We can break down performance spec into 3 areas: General Usage and Loading, Server Testing and Drive Throughput.
When booting the operating system and loading games and apps the Raid 0 array offers only a minor benefit in some instances over the single drive. From my past experience with load testing, the only time that significant results were achieved was when Sustained Throughput gained from an extra drive was coupled with quicker drive access time. Access time is associated with data location and/or partitioning setup, not to mention a faster rotational speed of the actual drive. Since our access time did not decrease tremendously from using small 25GB slices on the 7200.11, in most instances games booted 1 second faster, and the OS booted around 3 seconds faster when compared to the single drive. PCMark05 gave us an analysis of right around 2X faster than the single drive. However, PCMark05 is a synthetic benchmark that takes strict advantage of scaling a Raid 0 setup when computing a score, in essence doubling your score.
In our server testing portion using an 8k IOMeter benchmark, we had discovered that in a full random read web server configuration the two drive setup clearly has a huge advantage in performance by around 135%. When used in a database server with full random writing to the drive we experienced another incredible gain in IOP's of 120%. Normally, we would only be expecting a maximum theoretical increase in performance of 100%, but I believe by decreasing the Raid 0 volume to 50GB (2X25GB) from our original review and by installing the latest Matrix Drivers in Vista we were able to obtain excellent IOMeter results for a Raid 0 server setup.
Our final area of analysis will be drive throughput, video/graphic editing and working with larger files. When compared to the single Barracuda, the Raid 0 array of two drives proved to display a large benefit in many areas. Including cutting down large software installation time by 40%, cutting down large file transfer time by 60%, and finally increasing "true IOMeter measured" sustained read/write throughput on the drives by right around 90%. In the real world what does this mean? For task specific user's like a video editing studio this will essentially double your performance when working with larger files in a pure editing environment. To the average Joe, the Raid 0 array will increase file system performance including decreasing software installation time and dealing with your large collection of photos, MP3's or compressed video files. You will experience a slight improved benefit in most of your file operations, windows will boot a tad faster, and you may or may not experience quicker game/app loading only if the software has a buildup of larger files to read and can take advantage of the array.
E-Mail the Author