My Heavy Utilization Reserved Medium-sized SQL Server Standard instance on AWS is currently running to the tune of $0.279 per instance hour, after the upfront $1,620.00 not including EBS storage and snapshots (bandwidth is a non-issue as this is only accessible to other AWS instances).
I had no choice (and it made sense) but to purchase a reserved instance given what Amazon is charging for standard SQL Server instances ($0.744 per/instance hour), well, that and in any case I needed it running 24/7.
Considering I am tied to SQL Server I need to look at some alternatives to AWS, it was a nice way to bridge a gap and quickly migrate after acquiring a new application but when the year is up, the viability of this AWS product is going to be scrutinized.
Another addition for AWS was recently released and adds a nice feature to the Route53/S3 combo, the ability to failover to an S3 bucket if your primary site goes down. This can be paired up with both weighted and latency based routing and can be configured whichever way is best for you.
What’s more intriguing about this is not the ability to failover to an S3 bucket but the ability to failover to another region, Say if us-east-1 goes down, redirect the traffic to us-west-1 or maybe a backup on a completely different provider…say Rackspace or even your own private cloud to save costs provided you can handle the traffic.
Then again, does AWS ever go down? Ha, that being said, this sure beats my website being load balanced over multiple Micro AWS instances…
Today Amazon released another big enhancement for EC2 users, CloudWatch alarms can now be used to detect idle/unused or even underutilized instances and shut them down. Users have the option of first sending an email, shutting down the instance and/or even terminating the instance after a predefined period of time.
What does this mean? It means we can stop wasting money on instances that have been forgotten for a couple weeks, it’s also a great tool to use for keeping developer instances in check, the early email warnings provide an opportunity to prevent accidental terminations without taxing managers to monitor usage daily.