Amazon implemented a new predictive scaling feature for Auto Scaled EC2 (Amazon Elastic Compute Cloud) instances yesterday which uses machine learning to make predictions about traffic and automatically scale cloud computing instances accordingly.
“Using data collected from your actual EC2 usage and further informed by billions of data points drawn from our own observations, we use well-trained Machine Learning models to predict your expected traffic (and EC2 usage) including daily and weekly patterns,” Jeff Barr, chief evangelist at Amazon Web Services, wrote in a post on the AWS Developer blog.
Barr wrote that the machine learning model only needs one day of training to begin predictions and makes a new prediction every 24 hours of the next 48 hours. The predictions are based on user-defined configurations of scaling method, metrics tracked, and resource optimization.
“Once your new scaling plan is in action, you will be able to scale proactively, ahead of daily and weekly peaks,” Barr wrote. “This will improve the overall user experience for your site or business, and it can also help you to avoid over-provisioning, which will reduce your EC2 costs.”
Barr’s post breaks down the configuration of predictive scaling in Auto Scaled EC2 in detail.