Amazon Starts 2016 With Three Price Cuts

Amazon Web Services diminishes its own prospects of handsome quarterly profits through a set of price reductions. These cuts also make life harder for competitors.
Twitter's Top Data Science, Analytics, And BI Feeds
Twitter's Top Data Science, Analytics, And BI Feeds
(Click image for larger view and slideshow.)

Amazon Web Services started the new year with a round of price reductions on popular EC2 instance types. The C4, M4, and R3 were all reduced 5%, effective Jan. 1, for On-demand, Reserved, and Dedicated host instances.

The price reductions applied to AWS GovCloud as well as its most popular commercial installations: US East (Northern Virginia), US West (Northern California and Oregon), Europe (Dublin and Frankfurt), and Asia Pacific (Tokyo, Singapore, and Sydney)

Amazon, already the largest cloud supplier, appears ready to continue the competition with the likes of Microsoft, Google, and IBM, each of which would like to enlarge its own base of cloud customers. In some reports, Microsoft and IBM are growing faster than AWS percentage-wise but are starting from much smaller customer bases.

The last time Amazon lost the lead in cloud service price reductions was March 2014, when Google led the charge, matched by Amazon Web Services and Microsoft a few days later. Since then, neither Google nor Microsoft, despite their deep profits, have provoked another round of cuts by leading the reductions themselves. Both try to stay in step, however, with Amazon pricing.

If Amazon intends to continue cutting prices through 2016 on either additional instance types or its most frequently used instance types it's bad news for competitors. It means that Amazon isn't feeling the pain that markets sometimes inflict on companies that show slender or nonexistent profits. Amazon's cloud unit, AWS, is profitable, but the company as a whole averages a little better than break-even as it continues to expand and make heavy investments for the future.

The discounted instances include the general purpose M4s, which are hosted on modern servers using Intel Haswell CPUs, also known as Xeon 2.4 GHz E5-2676 v3 processors, according to a blog by Amazon chief evangelist Jeff Barr. The M4s come in Large, XLarge, double XLarge, quadruple XLarge, and 10-times XLarge instance sizes, according to the AWS instance chart.

They represent Amazon's general purpose balance of compute, memory, and storage. The M4 Large, for example, comes with two virtual CPUs, 8GB of memory, as much Elastic Block Store storage as necessary, and 450 Mbps in EBS network throughput. The 10XLarge instance comes with 40 CPUs, 160GB of memory, as much EBS storage as needed, and 4 Gbps in EBS network throughput. The M4 Large is now priced at 12 cents per hour and the 10XLarge at $2.394 per hour, according to the AWS pricing table.

Also reduced were compute-intensive C4 instances based on Intel Xeon E5-2666 v3 processors (also Haswells). C4 virtual servers come in large, extra-large, double extra-large, quadruple extra-large, and eight times extra-large sizes. The C4 Large has two virtual CPUs, 3.75GB of memory, and 500 Mbps of EBS network throughput. The C4 8XLarge instance has 36 virtual CPUs, 60GB of memory, and 4 Gbps of network throughput.

The on-demand C4 Large is now priced at 10.5 cents an hour, and the C4 8XLarge at $1.675 per hour, according to the pricing chart. (Amazon hasn't yet updated its API-based pricing information, which may be retrieved programmatically, but Barr said it will do so later this month.)

The R3 instances also come in the large through 8XLarge instance sizes. R3s are typically used for memory-intensive applications, such as large database and caching systems or in-memory analytics. They have the lowest cost per GB of memory of any EC2 instance type, and come provisioned with a hefty measure of solid-state storage. The R3 Large, for example, has two virtual CPUs, 15.25GB of memory, and 32GB of SSD storage. The R3 8XLarge has 32 virtual CPUs, 244GB of memory, and 640GB of storage in the form of two 320GB SSDs.

[Want to learn more about how Amazon leads in cloud pricing? See Amazon Counters Google Cloud Price Cuts.]

With the 5% reduction in price the on-demand R3 Large is priced at 16.6 cents per hour. The R3 8XLarge is priced at $2.66 per hour, according to the pricing chart.

"During the month, your billing estimates may not reflect the reduced prices. They will be reflected in the statement at the end of the month," advised Barr's blog.

The 5% reductions apply to Linux workloads that don't require a license, such as Ubuntu or Debian-based applications, Amazon spokesmen clarified in a follow-up communication to Barr's blog. Smaller reductions apply to Red Hat Enterprise Linux, SUSE Linux Enterprise Server and Windows workloads, Amazon spokesmen said.

[Editor's note: This article was updated to reflect Amazon's clarification regarding Linux workload pricing, and to emphasize EBS network throughput where necessary.]

Elite 100 2016: DEADLINE EXTENDED TO JAN. 15, 2016 There's still time to be a part of the prestigious InformationWeek Elite 100! Submit your company's application by Jan. 15, 2016. You'll find instructions and a submission form here: InformationWeek's Elite 100 2016.

Editor's Choice
Sara Peters, Editor-in-Chief, InformationWeek / Network Computing
John Edwards, Technology Journalist & Author
John Edwards, Technology Journalist & Author
James M. Connolly, Contributing Editor and Writer