Wednesday, June 11, 2008

Re: Eucaliptus


On 11-Jun-08, at 1:15 PM, Greg Pfister wrote:

Do you really want to package what you run as an (opaque) virtual
machine, or as some other format that the cloud infrastructure has
more insight into? How do you know what to replicate when horizontally
scaling when all you get is a pile of virtual machine images?


That's the tough question!

If one's going to automate and improve the sorry state of IT management & provisioning, it's going to need to involve exposing the architectural elements (and corresponding viewpoints) to be able to get a handle on stuff that goes beyond just "install this software, twist this knob".   Otherwise we're not really solving any problems, we're just automating one-off use cases at a time.    

As to where the current trend is progressing, for example,  I think the RightScale guys do have a good start on instrumenting the "VM image" up to something richer like, say an auto-scaling MySQL cluster (see  http://wiki.rightscale.com/ for screen shots)...

On the other hand, all the options can be a bit overwhelming.   I don't see an exposure of the architecture-level assumptions -- it's targeted more at a seasoned administrator in assisting the provisioning of some 'pre-baked recipes' like LAMP stacks.

Cheers
Stu Charlton
Elastra


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Cloud Computing" group.
To post to this group, send email to cloud-computing@googlegroups.com
To unsubscribe from this group, send email to cloud-computing-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.ca/group/cloud-computing?hl=en
-~----------~----~----~----~------~----~------~--~---

No comments: