This switch is well built with a metal case, performs well, and is economical. I did lots of benchmarks and found that I got maximum speeds while doing multiple concurrent large inter-computer file transfers. I repeated some of the tests while bypassing the switch to see if there was a change in speed and there was none, so this switch kept up with whatever I could throw at it. If you want more speed, chances are you will gain more with better cables/plugs/jacks/etc. rather than looking for a faster switch.
I also used a Kill-A-Watt meter to see how much energy this uses. The amount was so low that it took days to get meaningful readings, and this will cost pennies a year and electricity is not cheap where I live. Also, the switch is economically priced. This switch caused my network diagnostic to misreport cable quality as poor, but although some utilities might give you false readings, your connection will not suffer.
I read elsewhere that this switch is not compatible with WOL (wake on LAN/magic packet) and the person who posted that got the information from Trendnet support. It turns out that this works PERFECTLY with WOL as long as your computer's port is set to auto negotiation. In other words, if there's any limitation, it's with the quality of tech support, not with this switch.
The only minor negatives are that the printing next to the LEDs is not on both sides, so it will appear upside-down if this is wall mounted, making it awkward to keep track of port numbers. Also, the slots for wall mounting could have been bigger to accommodate larger screws. The documentation is minimal, but this is pretty much plug and play.
UPDATE Jan 2011. I have read comments from customers who have had problems using multiple switches of this model on the same network. I have three of them. One is connected with approximately 25 meters of Cat-5e cable to another on a different floor in my home. The cable run is UTP, about half vertical, and half horizontal though an attic with no shielding (except distance) from other cables such as power cables. A third switch is connected with approximately 8 inches of the same type cable, allowing me to have more ports in that location. Benchmarks of file transfers will not give a true picture, since the weakest link (hard drive speed, overhead from data encapsulation, switches, network cards, wall jacks, drivers, cable, etc.) will be the limiting factor. All other parts of the chain will perform at least as well as the benchmark suggests.
I tried copying about 450 GB of relatively large and barely compressible data coming from images of blu-ray and DVD sources. Transferring through all three switches, full length of the cable, etc. gave me transfer speeds of about 96MB/sec. Presumably, the hard disk arrays are the limiting factors in my system, but the switches clearly had no trouble with those sustained speeds. I used Windows 7 drag and drop over a network, which is not among the most efficient methods. A series of smaller files such as JPEGs gave me "speeds" closer to 15MB/sec, which really means that the speed over the wire was probably about the same, but a large amount of the data was overhead encapsulation.
Although it's likely that the switch was not the limiting factor, it's clear that users who are getting poorer performance have other factors that limit performance, making it impossible to say that the switches are the cause of their performance issues. A direct run of cable from one of the computers to the other, bypassing all three switches, would show if there were any increase in performance, thus showing if the switches themselves (or something related to plugs, cables, or jacks) caused the problems. But the speeds I got from several such tests were consistent, and higher than what I would have expected given that I used Windows drag and drop over a network.