Amazon had a unknown server outage today starting at around 5am and lasting until quarter-past 7am. The reports in main stream media and blogs have been comical based on how quickly people want to throw AWS (Amazon Web Services) under the bus. Granted Amazon does not have a SLA (service level agreement) but, is that really any different than owning your own infrastructure. Unless you are hosting applications on the edge (think Akamai) and other CDN’s, hosting is bound to fail at some point….even SLA’s that offer 99.999% uptime have cryptic legal language that exonorates them from legal damages. I had a webhosting company for 2 years before selling off the assets and referring business to parnters and downtime is inevitable. It’s how well you manage the downtime and your customer service requests that matter.
Startup Addict has AWS in our business plan…although not implemented for 1.0 and 2.0 release it will be incorporated by 2.5 release, it just makes good business sense. The impressive aspect of the downtime was how quickly it came back up and you better believe Amazon will be throwing resources at rectifying and preventing this problem in the future. As I reported on earlier, numerous startups are attempting to compete with AWS and are offering SLA’s but, this practice is limited because the SLA is inevitable for Amazon. The reason one is not in place now is very clear after this event –Amazon is still working out the kinks. A new startup will not find a better economy-of-scale anywhere than at Amazon web services for infrastructure (unless your VC funded and have a high burn-rate).