For the 19th time, Amazon Web Services (AWS) dropped prices for its EC2, RDS, ElastiCache and Elastic Map Reduce offerings. And AWS is not alone in making pricing changes; during the first week of March 2012, major price cuts were also made by Google and Microsoft.
Why the major shift in market pricing now?
Our take is that increased competition and a growing awareness among users of the challenges in controlling costs within the cloud are both factors in the rush to reduce the fees. AWS continues to push down reservation fees to stay ahead of an increasingly crowded marketplace – and others are taking their cues from Amazon. These reduced fees may save businesses money, but only if they can get clarity of how their business objectives shape cloud requirements and can match those needs to the appropriate pricing model for the cloud.
What the Price Changes Mean for Cloud Costs for AWS
We evaluated and implemented new recommendations for customers within less than a week of the AWS pricing cuts. These recommendations are already being implemented by end users to continue to cut overall spending/over-provisioning within the AWS cloud.
With 150,000 virtual instances managed daily for our customers, we were able to see some clear trends in how the AWS pricing will impact savings.
We found that the price change on average was 10% per AWS EC2 customer– rather than the 37% touted by Amazon in its announcement.
The reasons behind this were:
- Main impact was to reserved instances, but the majority of AWS instances are on-demand.
- On-demand prices for some instances (the high-memory family: m2.xlarge, m2.2xlarge & m2.4xlarge) were reduced by at least 10%, but the majority of instances in AWS are m1.small and m1.large, and the price reduction for these instance types is ~5.3%.
- About 40% of AWS instances are running Windows. The price reduction for Windows platforms was much less dramatic
One clear benefit we discovered with the new AWS pricing models is that before the reductions in prices, AWS Light reservations were economical only if production was in place for more than four months per year (~33% of time), so there was no possibility of using the instances in short-term planning. Now, in typical cases, Light reservations start being cost-effective after only ~2.5 months of use. This change enables users to have a lower-cost reservation solution that can be utilized in short-term planning.
If you are interested in seeing how these price changes will impact your organization’s potential savings, register with Cloudyn https://app.cloudyn.com/register.html for the free trial.