One of the traditions of BrainShare is the Friday keynote, where Novell shows off some of the things that are in the works. This year was no exception (video here). Many of the demos were results of Novell's "hack week," where people were encouraged to contribute to any open source project, or even start a new one.
The demos started with updates to the Xen Virtualization project. Soon you will be able to boot a server off a USB drive containing the Xen Hypervisor. No need to install an OS, just boot up and its ready to run Xen Virtualization images. Combined with Novell's Zen Orchestrator product, the Xen appliance self-registers with Orchestrator and is ready to accept virtualization work.
A number of demos from Novell's Hack week had to do with integration of the Evolution program with other projects. For example, one was a plug-in to allow Evolution to work with Novell's Teaming + Conferencing product.
Two other engineers created a new project called Tasque. This is a desktop application to the "remember the milk" Web task-management site. They had it up and running early in the week, so they added integration with Tomboy notes, and support for a project from last year's hack week (giver) to send out tasks to others over the network.
Future development work on the Novell Teaming + Conferencing product was shown. Billed as a solution to help with information overload, after the demo I'm not really convinced it would help -- I figured it would make things worse (but perhaps a bit more organized).
Green computing was one more of the buzzwords for the week. Another demo was geared toward how to reduce power and cooling costs of the data center. For example, using PlateSpin's PowerRecon to analyze your data center workloads, you can use ZEN Orchestrator to schedule and move virtualized workloads around to fewer physical machines. The example used was for a company that has high server demand in the day, but low use overnight, ZEN Orchestrator can move the virtualized workloads over to just a few physical machines and turn off the unused servers. Then as the anticipated workload goes up, the machines can be turned back on and the load distributed back out.