We need now to talk about performance. As you may imagine, you can’t expect a remote file server to deliver files as fast as your local hard disk drives, especially because it is accessed thru a network cable, which provides far lower bandwidth than the local SATA cable (Gigabit Ethernet provides a maximum theoretical transfer rate of 125 MB/s while SATA-300 drives provide a maximum theoretical transfer rate of 300 MB/s).

The single most important thing you must have to do in order to improve performance of any NAS box is to upgrade your network to Gigabit, as we explained on the first page of this review. If you don’t do that, you will access your NAS box only up to 12.5 MB/s. We obviously used a Gigabit switch during our tests.

We made a lot of measurements with a program called ATTO Disk Benchmark (version 2.36), which is available inside a package called “ATTO HBA Utilities” from ATTO. This program offers two tests, one called “I/O Comparison”, where it writes files on the disk being tested and then reads them back, and one called “Overlapped I/O”, which performs the same task but allowing these tasks to be divided and executed in parallel (“queued”, in the industry jargon). By default the program divides each task into four parallel operations and we kept this value. This program also transmits data using several different block sizes, from 512 bytes to 8 MB to show the difference in performance the block size makes.

We had three goals with our benchmarking. One, more obvious, was to compare the performance of the reviewed NAS box to other NAS solutions. For that we installed two Seagate Barracuda 7200.10 160 GB hard disk drives and configured them in RAID0 on each unit. The second goal was to compare the performance of the reviewed NAS solution to the performance of a regular PC sharing the same hard disk drive (Seagate Barracuda 7200.10 160 GB installed as a single drive and also two of them installed as a RAID0 array), in order to simulate a typical home/office environment and see if the reviewed unit provides any performance advantage of simply sharing your PC hard drive. And the third aspect was comparing the results with the performance of the hard disk drives installed locally, i.e. on the same PC we were using, to see the drop in performance when you read files located on another place compared if they were installed on your local hard disk drive.

On the next two pages we are comparing below the results achieved with 128 KB blocks. We will present the complete results for all block sizes after these pages. The PC that was sharing the hard disk drive was based on an ASUS P5K-E motherboard (Intel P35 chipset, so RAID was controlled by ICH9R) with Intel Core 2 Duo E6600 CPU, 2 GB RAM (DDR2-1066 Dominator from Corsair) and GeForce 8800 GTS 320 MB from Gigabyte.