Yes. We had rolled out a Devpay-backed product called "Cloud Gear" about a year ago and ended up pulling it off the shelf due primarily to two irreconcilable issues:
1. their accounting process appeared to have some fundamental innacuracies at the time. We tracked the launching of instances and our numbers did not match up with theirs. There was anomalous usage behavior and Amazon eventually admitted there had been tracking glitches on their end - difficult to deliver a service if the underlying accounting can't be trusted. Note: this may have been (likely has been) remedied since but it was problematic around January '10.
2. As another commenter noted, the signup process for a newcomer was too much to ask of someone who didn't already happen to have an AWS account. And even for those who did it was still a cumbersome hamster maze to put them through and undoubtedly hurt the adoption numbers.
#1 has likely been resolved. #2 is inherent to this type of on-behalf-of arrangement. It is an interesting model and could definitely enable some innovative services. We just found it not ready for primetime when we had tested it back in late '09-early'10.
We tried a while ago. My suggestion: only use it if you're specifically selling to Amazon users (e.g. a EC2 image for Solr). If you're just selling a general SaaS or server product, it's hard enough to get someone to sign up for your service; don't make them sign up for another service too.
Comments
Has anyone used this?
Yes. We had rolled out a Devpay-backed product called "Cloud Gear" about a year ago and ended up pulling it off the shelf due primarily to two irreconcilable issues: 1. their accounting process appeared to have some fundamental innacuracies at the time. We tracked the launching of instances and our numbers did not match up with theirs. There was anomalous usage behavior and Amazon eventually admitted there had been tracking glitches on their end - difficult to deliver a service if the underlying accounting can't be trusted. Note: this may have been (likely has been) remedied since but it was problematic around January '10. 2. As another commenter noted, the signup process for a newcomer was too much to ask of someone who didn't already happen to have an AWS account. And even for those who did it was still a cumbersome hamster maze to put them through and undoubtedly hurt the adoption numbers.
#1 has likely been resolved. #2 is inherent to this type of on-behalf-of arrangement. It is an interesting model and could definitely enable some innovative services. We just found it not ready for primetime when we had tested it back in late '09-early'10.
We tried a while ago. My suggestion: only use it if you're specifically selling to Amazon users (e.g. a EC2 image for Solr). If you're just selling a general SaaS or server product, it's hard enough to get someone to sign up for your service; don't make them sign up for another service too.