News

Cloud Innovation Focusing on Temporary Jobs

• By David Ramel
• 05/21/2015

Call them Spot Instances, Preemptible VMs or whatever, but short-lived, non-mission-critical, temporary cloud compute jobs are emerging as a second front in the cloud compute wars.

The big cloud players can only cut prices so much, and brand-new, major innovations are few and far between. Thus providers such as Amazon Web Services Inc. (AWS) and Google Cloud Platform are finding new ways to compete, such as offering unused capacity for ephemeral compute jobs.

Google this week announced Google Compute Engine Preemptible Virtual Machines, a beta program that provides regular compute instances at a lower price "except for one key difference -- they may be shut down at any time," in a blog post. The Preemptible VMs will only run 24 hours max and can be shut down with only a 30-second warning.

"While that may sound disruptive, it actually makes them a great choice for distributed, fault-tolerant workloads that do not require continuous availability of any single instance," Google exec Paul Nash said. "By not guaranteeing indefinite uptime, we are able to offer them at a substantial discount to normal instances." That discount results in fixed pricing that starts as low as $0.01 per hour.

While industry pundits contemplated the usual tit-for-tat response from primary cloud rivals AWS and Microsoft Azure, AWS had already laid the groundwork for improving its comparable product, Spot Instances, with the acquisition of Spot Instance bidding specialist ClusterK earlier this month.

And, on the same exact day as Google's announcement, AWS took to its own blog to announce a new tool to simplify the management of Spot Instances in fleets that can number in the thousands.

Although AWS has been quiet about the ClusterK acquisition, it did confirm the move to buy the company that helps customers save money by managing the bidding process to find the best deals on Spot Instances, whose prices can constantly vary, unlike their Google counterparts.

"Spot Instances allow you to name your own price for Amazon EC2 computing capacity," the AWS site states. "You simply bid on spare Amazon EC2 instances and run them whenever your bid exceeds the current Spot Price, which varies in real-time based on supply and demand. The Spot Instance pricing model complements the On-Demand and Reserved Instance pricing models, often providing the most cost-effective computing capacity, depending on your application."

ClusterK uses proprietary technology to leverage the bidding process to find the best deals with a special algorithm and to guarantee uptime so mission-critical jobs can run on the Spot Instances.

With this technology now in house to improve the management of the bidding process to secure Spot Instances, AWS this week announced a new tool for managing the deployment of multiple Spot Instances.

Amazon EC2 Spot Fleet API lets users manage thousands of Spot Instances with one request, AWS exec Jeff Barr said in a blog post.

"Today we are making EC2's Spot Instance model even more useful with the addition of a new API that allows you to launch and manage an entire fleet of Spot Instances with one request," Barr said. "A fleet is a collection of Spot Instances that are all working together as part of a distributed application. A fleet could be a batch processing job, a Hadoop workflow, or an HPC grid computing job."

Barr said that AWS customers previously have been writing their own custom code to manage the deployment of these fleets, something that is no longer necessary.

"Instead, a single API function (RequestSpotFleet) does all of the work on your behalf. You simply specify the fleet's target capacity, a bid price per hour, and tell Spot what instance types you would like to launch. Spot will find the lowest priced spare EC2 capacity available, and work to achieve and maintain the fleet's target capacity. One call does it all, as they say ...."

Barr said this was an initial release of a new product, so much more functionality is planned, such as a weighting system "to express the relative power of each of your launch specifications in numeric form." He didn't say if any ClusterK technology will be used in the new initiative.

Meanwhile, Microsoft has been relatively quiet on this front, though it has vowed to match any price cuts by AWS. The "big three" cloud providers have a history of matching price cuts such as those announced this week by Google in conjunction with its Preemptible VMs. Which leaves industry pundits to ponder the next retaliatory move coming out of Redmond.