Google announced major price cuts for some of their cloud services including Compute Engine, Storage and BigQuery. But their price advantage lasted for only for 24 hours because the following day Amazon slashed the prices for some of their services – EC2, S3, RDS, ElasticCache and MapReduce.
Following is a summary of AWS’ price cuts:
| Service | Price Cut by % | Comment |
| EC2 – on demand | 7-40 | Price cuts depend on instance type. |
| EC2 – reserved | 10-40 | Price cuts depend on instance type. AWS has also volume discounts resulting in discounts up to 68%. |
| S3 | 36-65 | An average of 51%. |
| RDS | 28 | Average. |
| ElasticCache | 34 | Average |
| MapReduce | 27-61 | hs1.8xlarge now costs <$1,000/TB/year. |
Jeff Bezos Barr, Chief Evangelist for AWS, is the one who posted the price reductions on their website, reminding customers that this is Amazon’s 42nd cut since 2008. But, commenting on his blog post, some of the readers thanked Google for “reducing our bills.”
RightScale compared AWS’ new prices against Google’s, concluding that they are exactly the same for on-demand standard instances, but Google still has cheaper high-memory and high-cpu instances by 33% and 16% respectively, but AWS high-memory provides 30% more memory and AWS high-cpu has double memory and a SSD, so their offer is actually roughly the same.
When it comes to 1-year reserved instances vs. sustained-use, AWS’ standard and high-memory instances are 10 to 20% cheaper than Google’s 100%-use, while high-cpu ones are more expensive (2-3%). AWS has a major price difference for 3-year reserved instances which are 32-48% cheaper than Google sustained-use. But there is a difference in their models too: one has to pay up-front for AWS reserved instances while Google sustained-use price is calculate as-you-go. RightScale outlined three price advantage scenarios for different types of cloud users:
- For cloud users who have plenty of cash and can accurately predict their usage, AWS RIs can save money.
- For the many cloud users who are still growing their cloud workloads quickly and leveraging a blend of on-demand and reserved instances, the most cost-effective option will depend on the exact type of instances required and the ability to make effective use of RIs.
- For cloud users who don’t have a lot of predictability in their workloads and are using primarily on-demand, they will likely save money with Google.
Microsoft pledged their “commitment to match Amazon Web Services prices for commodity services such as compute, storage and bandwidth,” a year ago when announcing price cuts for some of their Azure services. We kept this article on hold hoping that Microsoft would make a price announcement this week, but nothing came out of Redmond by the time of its publishing.