Wow, I'm not sure it was the brightest idea to try to blog on each demo here. There's a lot of them, and six minutes isn't enough time, at least for me, to form a meaningful opinion about most of these companies. Here's a little recap after some further investigation.Like most people, what I learn from these demos is which of these companies I don't need to visit. As I mentioned earlier, I found almost none of these companies to be laughable, which is a real improvement over the spring show. There's also not much enterprise-class technology being shown here, though there is some. That, too, makes sense to me. A good consumer technology should be obvious and simple, but good enterprise technologies are often all about nuance and flexibility, something that's very hard to get across in a six minute presentation.

I visited a few of the vendors to learn more about their products.

Digital Fountain was the first demo of the day. It showed its content delivery network streaming flawless broadcast-quality video even with a 20% loss rate. It does that by using forward error correction techniques and by pulling content from potentially dozens of servers at a time rather than just one. The first application for the technology has been in licensing it for use in private networks, some of them pretty big ones. This CDN is a pay-for service that may have a slightly harder time catching on since it requires support at the client end. Still, the technology seems to work as promised.

Global Communications followed Digital Fountain in the morning. The company is basically looking to use satellite links, point-to-point free space optical links, and any copper it can find in the last mile to deliver content to places were it couldn't otherwise be delivered. One of the first applications is bringing various services to third-world countries. I'm not sure Global Communications is in the same class as most of these startups, as it's already delivering these services to many parts of the world.

Many of the infrastructure technologies were interesting, if hard to demonstrate. Qumranet, FusionIO, and Phreesia all had interesting products that should have real-world applications, though FusionIO may have to worry about the likes of SanDisk jumping into their market.