As far as testing is concerned, you cannot run the same battery of tests on a networked device as you would a physical drive that is connected directly to your machine. Given this limitation, my main goal was to determine transfer speeds as best I could and to do large file transfers to make sure that the drive was able to read and write data accurately without errors or corruption.
The second task was the easy one. By using a simple Windows feature of copy/paste, I transferred large amounts of data from my primary hard drive to the network share and then ran a MD5 hashing on the copied files to ensure an accurate transfer. In all instances, there was no data corruption of any of the files. I used single large files, multiple large files and multiple small files in these tests and never had a hitch. So reliability looks to be proven with no issues of concern.
Testing data transfer rates is a bit more difficult. While Sandra can still be used to test a network drive, HD Tach refuses to notice this type of device. So we'll rely on a Sandra run and timed transfers of an approx 500MB file. I don't expect stellar speeds from a network share, but I'd like to see something that is at least comparable to what we should expect from this type of connection.
Sandra is a purely synthetic benchmark, but since we're looking at data transfers it has merit. Running on a TCP/IP protocol with a 100TX switch, we see an overall rating that is a little under what their baseline is. This isn't too surprising since there are three active machines working over the network all the time. I did not single out just this connection since I wanted to give an accurate assessment of what you could expect from your own home network. With multiple machines in a household becoming more prominent all the time, this scenario isn't out of line with a normal network.
With transfer speeds ranging from 3100-3200 kB/sec, we won't set any speed records, but as I mentioned before this is a network device and this simply isn't a possibility. Given this situation, I was a little impressed with the 27ms access times. This is pretty good for a networked share.
When timing a data transfer, I use the demo file of Far Cry. It weighs in at just over 500MB and because of its large size it can give us a little information about how fast a drive is able to accept data flowing both toward it and away from it. The times shown above are very workable for the networked condition we're running.