How Cost Optimization Helps You Get the Best Value From AWS

AWS Elastic Container Service (ECS) fully manages orchestration to simplify deployment, management and scaling of containerized applications. By describing an application and the resources required, Amazon ECS launches, monitors and scales that application across flexible compute options with automatic integrations to other supporting AWS services.

“This can help to realize up to a 70 percent savings, compared to the on-demand cost of running stateless and fault-tolerant applications,” says CDW Canada Field Solution Architect Sylvie Sema. “It also makes it easy to manage and deploy containers running applications.”

AWS Trusted Advisor provides actionable recommendations based on usage, configurations and spend analysis then identifies ways to optimize your AWS infrastructure, improve security and performance, reduce costs and monitor service quotas. AWS Trusted Advisor sends alerts when service quota overages occur or when more than 80 percent of that quota is reached.

AWS Compute Optimizer analyzes the AWS resources your organization runs and generates possible downsizing or upsizing service recommendations by identifying computing or storage resources that may be over- or under-provisioned.

“It identifies resources that are optimal and not optimal – basically what you really need and don't need,” Sema says. “Then it offers resource recommendations to reduce costs, improve performance or a combination of both. Taking advantage of compute optimizing can lower your cost by up to 25 percent.”

AWS Budgets allows you to set an AWS service spend limit and manage costs and usage in order to adhere to that budget. You can continually monitor your budget status from a dashboard and add email or IM notifications or alerts if you near or exceed your spending thresholds.

“You can set custom cost and usage budgets to easily manage your AWS spend and then you can still monitor your budget status from the AWS budget overview,” Sema says. “You can track your costs or (manage) across multiple dimensions by adding filters related to your service and your accounts…and receive daily, weekly or monthly reports that allow you to monitor your AWS (service) performance.”

CDW Canada Senior Field Solution Architect, Ally Meer-Hossen, says all of these AWS services provide a good overview of a customer’s cloud environment. However, CDW brings added value by going further and looking more deeply into services designs and operations for our customers.

He highlights a recent example of an AWS customer that CDW supported during a migration from standalone virtual machines to an autoscaling group.

“There were couple of tweaks that needed to be done within the applications to support the autoscaling group, which meant these had to be decoupled with less dependencies,” he explains. “When we implement AWS optimization tools for a client, they can typically recover approximately $2,000 USD per month (in services costs). But when we looked at what this customer was running and going deeper to help them optimize, we were able to achieve a cost recovery of approximately $35,000 per month. 

“This is how CDW demarcates itself from the competition,” Meer-Hossen adds. “CDW can bring tremendous cost-recovery value to AWS.”

Tagging is another strategy for identifying potential cost reductions. The effort involves discovering if an organization can appropriately shut down certain AWS services when these are not being used during certain periods of time, such as on weekends or after business hours. It’s an important activity; research shows more than 40 percent of technical and business professionals use automated policies to shut down workloads after hours and to right size underutilized instances.

“Tagging gives you a facility to orchestrate,” Meer-Hossen says. “It is something that needs to be designed and purposed properly. CDW can help.”

Ongoing monitoring to discover usage trends is also important. Meer-Hossen says most organizations don’t monitor cloud computing activity and AWS cost optimizing tools do not look for usage trends. He suggests, for example, that examining data from equipment such as virtual machines and SQL servers can reveal usage trends that may suggest a benefit from downsizing services. He cites customers who often overprovision random access memory (RAM) in their service offerings without examining what they actually consume.

“What do they use on a daily basis? What are the peaks? That information is important,” Meer-Hossen says. “By monitoring usage, you might want to resize your resources to reduce your costs on a perpetual basis.”

Also look at your “burst trends and patterns” by examining data collected by AWS Cloud Watch and analyzing the metrics of each resource, he says. It is typically a manual effort, but scripts can be created to automate at least some of these processes.

“I have seen…that, if you just go heads down into cost optimization, it’s not going to help you a lot,” Meer-Hossen says. “It might bring down your costs by one or two thousand dollars a month, depending on the usage globally. But if you look at the trends – which means analyzing activity over a period of one to three months, to identify concise and precise trends – that data will significantly help to improve your cost optimization efforts.”