A Closer Look:

Now that I'm holding the enclosure in my hands, I can tell it's fairly light, (just a hair over a pound) and made mostly from an aluminum pressing. The shiny side is a glossy black perforated aluminum plate with a mirrored strip along the top, printed with the Ineo brand name - this comes off as very classy, also lining up nicely with the status display on the front panel. High marks for industrial design so far. Now it's time to pull out the key and take a look at this toolless drive insertion mechanism. Simple - you stick the key - really just a plastic peg, a bent paperclip would work too - into a small hole on the top of the drive (it really was made to be used vertically) and the front panel swings down on a hinge. Thanks to the wonders of the modern SATA connector, you just drop a drive in, ensuring it goes in the right way, and close the trapdoor. That's all there is to it. Drive removal is also a snap - poke the key into the hole, and the trapdoor opens, levering the drive out of the SATA connector. Pull straight out on the door, and a spring-loaded tray pulls the drive out far enough to grasp, sliding back into place as soon as you let go. The mechanism is very simple in concept, very clever in implementation, and a breeze to use. The only problem with this is that the case of the hard drive and the enclosure only come into contact at a very few points, minimizing thermal transfer and removal of waste heat. Unfortunately, in a case of fashion over function, this enclosure was not built with any sort of active cooling built in - not a fan in sight (or even cleverly concealed). This really is not an uncommon trait, but it still bears mentioning.

Testing:

We were fortunate to be able to give this enclosure a good long thorough test, and get a feel for the sort of heavy use scenarios that make the best tests. The procedure was simple - one of our testers set this up as a data drive on a personal computer, and used it as such for a month. At the end of that, here's what he had to say:

• I wish my motherboard supported eSATA - that's a really nice feature.
• Short USB cable is short. I wanted to put this up on my desk, but I was forced to leave it on top of the tower because the cable would not reach.
• During extended, thirty to fifty gigabyte data write operations, the data rate dropped off with time, as the drive throttled itself down in order to prevent itself from overheating. This was using a standard 7200 RPM drive - I don't want to know what a 10,000 RPM high performance drive would do in there. While there are four small divots on top of the drive to let you stack them stably, thermal concerns suggest that this would be unwise at best.

While I was not able to duplicate the thermal issues my friend had, I attempted to approximate this by benchmarking the sort of speeds one is likely to see on an extremely heavily loaded system. I should point out that this is not the peak rate the USB controller is capable of, but more accurately represents the sort of rates one would actually see when working with the drive enclosure in real-world use cases. The test system is a Summer 2007 Macbook Pro with a 2.4 GHz Core 2 Duo, 2 GB ram, and 800 MHz bus speed. The drive was connected directly to a system port with the included cable, and the system was placed under load - playing music, downloading software updates via BitTorrent, twenty six (!) web pages opened, and web authoring tools running under Rosetta were launched. Xbench was finally launched and pointed at a fresh HFS+ partition on a Western Digital Caviar SE 7500 RPM hard drive manufactured in June 2005. A standard disk test was executed, resulting in the following:

Uncached small-block random read compared favorably to even Firewire 800 drives, producing essentially equal transfer rates. The eSATA interface was not benchmarked, being a glorified pass-through for the drive itself, as frankly the USB controller is where all the on-board logic happens. Despite the abusive conditions it was tested under, the drive and enclosure held up fairly well, producing good transfer rates, even among the difficult random tests - where it actually exceeded the read rate of a similar Firewire 800 drive when confronted with 4K blocks. I've said it before, I'll say it again - USB excels when confronted with many small operations, though it bogs down on large files - and this seems to be a fine example of a USB controller done right.