AWS Auto Scaling simplifies management across services

AWS has simplified its auto scaling capabilities as it continues to pool its expansive product list.

AWS Auto Scaling helps reduce the effort it takes to manage applications that scale across services because it consolidates the controls for individual applications in one place. In the past, users had to manage products deployed through auto scaling, individually.

Rapid scalability has always been a selling point of public cloud usage. Users can build those capabilities themselves, but since 2009 they've also been able to use auto scaling to set policies and alerts to track and manage those tasks for Elastic Compute Cloud (EC2). Over time AWS added similar capabilities for containers, application streaming and several of its native database services, but they had to be managed separately, or required users to build custom scripts to coordinate the various programs.

With this latest update, AWS Auto Scaling, users can set policy standards across EC2 Auto Scaling groups, EC2 Spot Fleet, Amazon Elastic Container Service tasks, DynamoDB tables, DynamoDB global secondary indexes and Aurora replicas through either AWS CloudFormation stacks or AWS Elastic Beanstalk. Users can define their scaling policies to meet best practices for availability and cost (or a mix of both), and set custom parameters that will work across these services.

You still have to architect around the application [but] it will definitely make it easier to configure and save time.

"This won't suddenly make WordPress scale up to meet a global audience," said Jeff Aden, co-founder and executive vice president of strategic business development and marketing at 2nd Watch, an AWS managed service provider in Seattle. "You still have to architect around the application for that, but it will definitely make it easier to configure and save time."

Like most product releases from AWS, Auto Scaling has its early limitations. It is only available in five of Amazon's 18 global regions: U.S. East (Northern Virginia), U.S. East (Ohio), U.S. West (Oregon), EU (Ireland) and Asia Pacific (Singapore). AWS said it will add additional features and options based on customer demand, such as the ability to flag resources with more than just CloudFormation and Elastic Beanstalk.

There is no additional cost for the pooled service, beyond what customers already pay for resources and alerts.

PropertyRoom.com uses AWS' EC2 auto scaling features to prepare for big traffic spikes in some of its web tiers, but scaling across services is time-consuming, said Kevin Felichko, CTO at the Frederick, Md., company. He's particularly excited about the potential to pull more AWS tools into Auto Scaling -- for example, to reduce the time for a Lambda function to respond to a request.

"If you can pre-warm those and keep it going for those types of big events, it just makes you feel like you have a bit more control over what's going on," Felichko said.

AWS has made some effort to streamline and simplify their many products and services, the sheer number of which can create problems for customers as they try to determine which tools to use. AWS Auto Scaling illustrates this effort. Other examples include EC2 Systems Manager for configuration and automation, AWS Migration Hub to move workloads to the cloud, and AWS Glue to manage different data sources for analytics.

Trevor Jones is a senior news writer with SearchCloudComputing and SearchAWS. Contact him at [email protected].