The ability to create virtual environments is perhaps the most useful feature offered by 8.10 Server in an age when virtualization has become the key to everything.
The 'encrypted private directory' option also sounds very cool, although I haven't gotten a chance to play with it yet. Privacy is a big issue on many of the multi-user servers that I deal with, and an out-of-the-box solution for making sure that certain users' files are unreadable and untouchable in the event that a machine becomes physically compromised would help to alleviate security concerns.
It's clear that these features are not just cursory additions to an Ubuntu release whose meaningful enhancements I'd expected to find limited to the desktop edition. Rather, they send a clear message that the Ubuntu developers put a substantial amount of effort into making sure that Ubuntu Server 8.10 would be significantly better than 8.04.
A Better Ubuntu, But for Whom?
I can't help but wonder, however, how many people are really going to benefit from these improvements. Few serious system administrators are likely to use Ubuntu 8.10 on their servers, since it only comes with eighteen months of support, compared to 8.04's five years. Having to revamp a server every year-and-a-half in order to ensure that it continues to receive security updates is neither fun nor productive, and for most people, running a production-environment operating system that has seen its end-of-life is not acceptable.
Sure, you can in principle upgrade a server to the latest Ubuntu release easily and painlessly over the network, just as you would a desktop machine. But unless you can afford the several hours of downtime that might result from something weird happening during the upgrade, you're not likely to do it—at least, I certainly wouldn't, given my numerous bad experiences with distribution upgrades on my Ubuntu desktops.
Maybe I'm wrong. Maybe there are lots of people out there who are going to use Ubuntu 8.10 Server for serious deployments, despite the hassle of having to upgrade before April 2010. If so, I hope these people will speak up to correct me.
But if Ubuntu 8.10, and other in-between-LTS releases, really do see very little action on production servers, then it calls into question the wisdom of releasing a new Ubuntu Server Edition every six months in the first place. Perhaps it would make more sense to upgrade only the desktop version biannually, and coordinate the Server Edition with the LTS release cycle.
It's difficult to draw conclusions without seeing hard numbers regarding Ubuntu Server Edition use, and I don't know of any surveys that record the number of non-LTS Ubuntu servers vs. those running 6.06 or 8.04, the two releases that are supported longterm. And I suppose that the cool new features that made it into 8.10 Server Edition are not hurting anything. But I wonder nonetheless whom they're really for.