AWS gives businesses access to a vast array of enterprise-grade infrastructure, on-demand and with no capital outlay. From servers to machine learning services to many different data storage options, AWS cloud solutions can be the answer to your business’s infrastructure challenges. But this labyrinthine abundance is not without drawbacks: AWS cloud infrastructure can save a business a huge amount of money over the long-term, but only if users keep a tight rein on expenditure and utilization.
It is all too easy to run up big bills on AWS, as many companies new to the platform have discovered. Spinning up new AWS instances takes seconds, and without focus, monitoring, and planning, an AWS deployment can quickly outstrip a business’s budget.
Choose the right instances in the right sizes
The most common AWS cloud issues are to use too many servers, servers of the wrong size, or of the wrong type. AWS includes a multitude of instance options so that users can align their infrastructure with requirements, but over-estimating requirements leads to idle cloud infrastructure — infrastructure the business pays for whether it is used or not. For large cloud server deployments, over-sizing by even a small amount can be costly.
Amazon provides several different AWS cloud server types, including on-demand, spot, and reserved instances, each with a particular cost profile.
A common mistake is to use on-demand EC2 cloud servers when reserved or spot instances are more appropriate. For long-running jobs, reserved instances provide a substantial discount compared to on-demand instances. The inverse mistake is also common: paying for one or even three years in advance for reserved instances that aren’t fully utilized.
There is no one-size fits all solution to cloud server sizing: it depends on the needs of a business or project. AWS users should assess their requirements — or bring in an AWS expert to carry out an assessment — and monitor resource use throughout the project to ensure optimal utilization as requirements change.
If you aren’t using it, shut it down
Low-levels of server utilization are an industry-wide problem, but with on-demand infrastructure there is no reason to keep paying when infrastructure is idle or under-utilized. It is, however, not unusual for AWS cloud users to leave unused EC2 instances and storage volumes up and accruing costs.
To ensure that infrastructure is used efficiently, AWS users should make someone responsible for monitoring and reporting on cloud infrastructure utilization, and ensuring that it is aligned with the needs of the business.
In partnership with Eplexity, an Advanced Consulting Partner in the Amazon Web Services (AWS) Partner Network (APN), NYI offers Financial Optimizer, a service that helps businesses with out-of-control AWS costs to regain control of their infrastructure spending. To learn more, get in touch today.