We setup a nice Ubuntu cluster for a customer in the financial services world recently. They wanted something that was similar enough to what they knew, and was as close to painless for them to use as possible. Make it like a Ubuntu system.
And make it easy to manage. Real easy.
The issue was and is, pretty much none of the major cluster distros really support Ubuntu. A few have some hacks to enable some level of support.
Add to this that most cluster distros start with the concept of a compute node installation. Speaking to these users … nah … they really didn’t want that.
Actually I think most cloud folks would do away with that if they could. Installation time isn’t free. Costs you money. So why bother?
So we built it. From power button time, until new machine is ready to use, is less than 60 seconds. Inclusive of the special low latency 10GbE fabric (not a contradiction in terms here). And mounting local disks (set of SSDs set up in a RAID0 for high speed local access).
Whats nice (though not finished yet) is each new machine they get, they simply connect in, turn it on, and walk away.
This instant on is, honestly, what I think most cluster folks in the cloud … and elsewhere, really want. That, and easy management.
We’ve got the latter covered. Just like with Bright, you modify one instance and its pushed automatically. We have a few more features that are relevant for us (we use this for our storage clusters for example), that are probably relevant for a few others. Just have to finish a few neat things for clusters, VMs, and a coupla-other-things …
Hope to have this going for our storage cluster at SC. Well, the storage cluster will be here. But it should be instant-on.