Intel’s Intel’s Network Attached Storage Performance Toolkit, or NASPT if you would prefer, will conclude our look at performance.

The Intel NAS Performance Toolkit (NASPT) is a file system exerciser and analysis tool designed to enable direct comparison of network attached storage (NAS) performance. NASPT seeks to discern differences in user level performance when a given client PC uses different remote storage solutions. To that end, NASPT uses a set of real world workload traces gathered from typical digital home applications: HD video playback and record, office productivity applications, video rendering/content creation and more. NASPT reproduces the file system traffic observed in these traces onto whatever storage solution the user provides, records the system response, and reports a rich variety of performance information. While NASPT runs on a Windows XP* client, the target NAS device may run any operating system.

Intel’s NASPT offers a number of strengths. First and foremost: consistency. With INASPT’s built in traces, performance numbers resulting in various tests are easily emulated across various systems and gives insight into how a NAS system would perform under real world conditions. Unlike with localized storage, we are not interested in separating software environmental factors as by definition and function, NAS storage units are heavily dependent on system software, operating systems, network settings and so on. Of course, if we were to simply test the bare hard drive performance at the system level our resulting numbers would be significantly higher.

With the trace files, we can simulate specified hard drive activity down to the distribution of random and sequential reads/writes, the spread of data over individual platters, or locality, and a number of access characteristics.

For our tests, we will be looking at four tests: HD Video Playback, Content Creation, Office Productivity, Copy To NAS and Copy From NAS.

Let’s examine the results, starting with the HD Video trace. This test consists of reading a single file approximately 4800 MB in size and for the most part consists of sequential reading, meaning that all the pieces of the file are located physically near one another on the platters of the hard drive resulting in little arm movement to retrieve data.

Content creation is exactly what the name suggestions. This test simulates the creation of a video. A total of 99 files are used the test is primarily write driven (95% write operations) with up to 64kb block sizes and 39.1% sequential operations.

With office productivity, we are a wide array of file sizes and generally random access. While in comparison with sequential throughout, random access as shown by the Office Productivity test does look rather grim, however is quite typical of NAS units.

Finally, the Copy To NAS and Copy From NAS is simply the recorded throughput rate of either copy to or from the Synology DS209, as suggested. This test includes a blend of random and sequentially written data and best represents day to day performance.

In these benchmarks we finally see JBOD taking more of a backseat to RAID 0 and RAID 1. The results are still in a very narrow bandwidth and the numbers are quite amazing for a two drive NAS. I’ve seen 4 or 5 drive NAS units not come close to these figures. The TS-209 Pro II’s results were only half of the DS209.

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14