If you work with Amazon Web Services, you’ve probably heard of spot instances. Buying spot instances on AWS is becoming a more popular way to buy IaaS because of the savings that the model can provide an organization.
Spot instances assign a price per hour on a specific type of virtual machine in a specific data center or region. You can bid a specific amount and if your bid is met, you get the computing power at the price that you set. It’s a more affordable way to buy IaaS based on the current cloud economy of sorts. Amazon maintains a spot price portal that is available for inspection at anytime.
ClusterK Predicts AWS Spot Prices
ClusterK has formulated software that allows organizations to form an offensive approach to the fluctuating spot prices found on AWS.
The founders of ClusterK thought that if they could predict the spot price of AWS, they could create a software platform that allows organizations to take advantage of the lower priced IaaS offerings. ClusterK works by using its pricing prediction methodology to distribute computational capacity as needed. This allows ClusterK to offer a fault tolerant environment for those that adopt its platform.
You might be thinking, what happens if a particular service’s cost suddenly spikes? Founder Dmitry Pushkarev says, “So when those [price] spikes do occur, in most cases, they won’t affect our capacity” He adds, “We’ll try to rebalance that well ahead of time.”
In the past, developers who used spot instances for their environments may have needed to adjust code to ensure consistency. With ClusterK, no such thing needs to happen.
ClusterK mentions that it has a good working relationship with the AWS spot instance team. Founder Dmitry Pushkarev said that the spot instance team at Amazon are big fans of ClusterK’s software as ClusterK only helps spur the adoption of Amazon spot instances.