openstack team mailing list archive

[Chris Behrens <chris.behrens@xxxxxxxxxxxxx>]

On Jul 8, 2011, at 5:11 AM, George Reese wrote:

> I would just like to re-iterate that I think the entire UUID approach is flawed and issues like this are one of the key reasons why.

The only problem I'm aware of is that developers using the EC2 API are not adhering to the spec.  If everyone treated them as strings, as they are supposed to be, then we wouldn't have to have this discussion.

That said, I do have a particular problem with the current UUID implementation in that I wish some sort of unique zone identifier were a part of it.  Accompany that with some other changes to zones and we could have more efficient zone routing.  A side effect of that would be that it is less work to come up with an ID that, if truncated, would also be unique to EC2.  Taking Ed Leafe's approach, you could remove the recursive zone checks.

I'm not sure I'd vote for that route, though.  It implies we kludge UUID generation just for EC2 in the heart of nova,  which I think is completely wrong.  I'm pretty much with Vish on everything he's said so far in this thread.

- Chris

This email may include confidential information. If you received it in error, please delete it.