An updated tool to stop and start instances is yet another sign of how AWS pushes its own techniques on its customers.
To manage its infrastructure, AWS puts a premium on operational efficiency through automation, and it has shared those practices with customers over the years, with features for autoscaling and load balancing. This week's addition of an expanded instance-scheduling capability is the latest signal that AWS wants its customers to emulate its own practices.
AWS Instance Scheduler supplants EC2 Scheduler, which was added in 2016 to help configure custom start and stop times for Amazon Elastic Compute Cloud (EC2). AWS Instance Scheduler extends those capabilities to Amazon Relational Database Service (RDS), so customers can automatically turn on and off instances using that service, as well. For machines that only run during regular business hours, this tool could save customers up to 70%, compared with running them all day, according to AWS.
To use the feature, customers define a schedule with a DynamoDB table and tag the appropriate resources. Using a CloudFormation template, CloudWatch triggers an AWS Lambda function to start or stop the instance at the desired time.
On the default setting, AWS Instance Scheduler costs approximately $5 per month in Lambda functions. If CloudWatch is used in conjunction, there is an additional $0.90 charge per month, per schedule.
The cloud cost struggle is real
Many new customers, particularly those accustomed to their own data centers, struggle mightily to manage their cloud costs. On premises, the amount of time a prepurchased machine runs is largely irrelevant, save for electric costs, whereas cloud instances are charged for every minute of usage. That creates huge waste for companies that overprovision or simply fail to turn off their machines when they're not in use.
Deepak Mohan
Many large enterprises quickly see the value to offload some operational expenses in the cloud, but they are slower to recognize how higher-level services help them get the best return on their investments, said Deepak Mohan, an IDC analyst. The AWS Instance Scheduler and other recent additions, such as AWS Auto Scaling, indicate AWS is adding these capabilities at the same time customers begin to realize they need them.
"It's making it easier to use services in a more efficient way," he said. "We see enterprises starting to move in that direction, so that is where Amazon is placing its investments."
If I need an instance for 40 or 50 hours a month, I won't buy a Reserved Instance, but then I may forget to shut it off at night. So, this is a beautiful [option].
E.J. Brennanfreelance developer and AWS architect
It's also notable that RDS has been added as a service that can add this automation, Mohan said. Databases have become AWS' biggest area of growth, and the cloud vendor sees those tools as one of the three key building blocks for customers, along with EC2 and Simple Storage Service.
E.J. Brennan, a freelance developer and AWS architect in Massachusetts who works with large enterprises, tried out the AWS Instance Scheduler and said it worked well, minus a slight glitch that shut down instances running before the scheduler was installed.
Previously, a user could write custom scripts to schedule downtime, but that would have taken too long to justify. Having customizable features readily available in your account will save considerable time and money, Brennan said.
"If I need an instance for 40 or 50 hours a month, I won't buy a Reserved Instance, but then I may forget to shut it off at night. So, this is a beautiful [way to avoid those unnecessary charges]," he said.
Trevor Jones is a senior news writer with SearchCloudComputing and SearchAWS. Contact him at [email protected].
Deepak Mohan
