While hundreds of companies were impacted this morning by the outage in the Amazon AWS cloud SendShorty was not impacted at all. Granted our users were not able to post to FourSquare because the FourSquare service itself was not working but that wasn’t something we could do anything about.
It is amazing to see how fast cloud services have grown over the last few years, hundreds of startups rely on Amazon’s cloud services for their day to day operations and when Amazon’s east coast data center started having issues, this morning, a very significant number of these startups also started experiencing issues. It begs the question if the people responsible for the architecture and design at the impacted companies thought the architecture and design through or if they just choose the cheapest solution they could find! (and it wasn’t only startups impacted by this outage)
At SendShorty we don’t just rely on one server or one cluster of servers in one data center. Our service is build on a geographic redundant clusters of servers, dedicated hardware that is up to the task! It limits us slightly in our ability to scale up our service although we could quickly build additional clusters in the cloud if needed, we just wouldn’t rely on one cloud services site!
This doesn’t mean we never experience issues, we have had hardware malfunction and even whole data center sites go down or loose connectivity. We carefully monitor all our hardware, the services running on them and our connectivity and quickly take action in case of failures but we won’t be impacted at the scale some of companies were impacted today!
