This is a subject I have much interest in and something we are working on very hard at Blue Gecko; Amazon is positioning itself as an Oracle partner and leader in the Oracle/cloud space and I want to know anything I can about what other people are doing, which is why this presentation was not what I was hoping it would be.
The gentleman presenting (name withheld – look it up if you’re really interested) basically used an Oracle-provided Amazon AMI, fired up a database, and wrote up some PowerPoint slides.
To be fair, he was approaching Oracle EC2 from a development perspective, where environments are ad-hoc, truly disposable, and the only relevant cloud-based advantages are cost and deployment speed. Not that those aren’t key, they certainly are, but he didn’t discuss any relevant performance benchmarks, the advantages to snapshots using Elastic Block Storage, or much outside of simple, disposable use of a pre-built AMI.
He also seemed persistently skeptical about Amazon’s trustworthiness: No one really knows what components make up EC2; Amazon says they can’t see your data or log into your servers; Amazon tells us that S3 is replicated in three places.
But do we trust them?
Geez. I don’t know. Do you trust Rackspace, Savvis, or any hosting company? Do you trust the sysadmin you just hired and handed the root password? Man, the paranoia was selectivcely thick.
I was hoping to hear a discussion or see some demonstration of Oracle best practices, even if they were best according to just the presenter. How did he deal with instance persistence? Was there a rapid deployment strategy? How did he decide to use EBS to leverage snapshots and data persistence? I guess I was hoping for a more operational spin. He spent a lot of time on EC2 basics: How to sign up, generating keys, using putty, logging in, etc.