Amazon.com today said it would offer variable pricing for clients of its cloud computing service. The pricing, called Spot Instances, allows users to bid for access to unused compute time for as long as their bid exceeds the current spot price for an hour of computing. The introduction of dynamic pricing is something our guest columnist Joe Weinman predicted this weekend in his post discussing different ways to price services in the cloud. From the Amazon release:
To use Spot Instances, customers place a Spot Instance request, specifying the instance family, size and the region they desire as well as the number of Spot Instances they want to run and the maximum price they’re willing to pay per instance. If a customer’s maximum price exceeds the current Spot Price, the customer’s instances will run until they choose to terminate the instances or their maximum bid falls below the Spot Price (whichever is sooner). Like other Amazon EC2 instances, Spot Instances can be terminated when they are no longer needed. If the Spot Price goes above a user’s maximum bid and the instance is terminated by Amazon EC2, the user will not be charged for any partial hour of usage.
Amazon notes that Spot Instances is good for applications that can have flexible start and stop times such as image and video conversion and rendering, data processing, financial modeling and analysis, web crawling and load testing. The pricing will help Amazon rid itself of unused capacity, much like hotels drop prices or sell certain rooms to discount services in order to ensure that they are as full as possible each night.
As our computers get smarter and our bandwidth becomes faster and more ubiquitous, this type of dynamic pricing will be seen more and more often. In addition to the cloud, mobile operators are eying dynamic pricing models, as are utilities as they implement smart grids. Fixed price may soon become a quaint anachronism, along with newspapers and landlines.