Rein in Runaway Costs

by Kevin McCaney

Runaway spending has become one of the most common problems in cloud infrastructures. As enterprises increasingly move operations away from on-premises data centers to take advantage of the cloud’s many benefits, the unexpected and unseen costs of migrating to and operating in the cloud can push systems over their allocated budgets.

Though cloud computing provides many efficiencies, there are several reasons why spending can exceed plans.

Developers will develop: Developers tend to focus on improving service and adding capabilities, rather than worrying about costs. Especially in an agile, continuous integrations/continuous delivery (CI/CD) environment, that kind of innovation can take a program’s costs beyond its allocated budget.

The downside of scalability: One of the strengths of cloud computing is its nearly infinite scalability — but that can also add unexpected costs to a project. In situations like a pandemic, the demand for data from other agencies and the public can grow quickly. The cloud makes it simple to meet that demand, but not without a cost. “The cloud is so agile, elastic in nature, that if the demand keeps coming, you just keep on spending the resources,” said Venkatesh Krishnaiah, Varcons’s director and founder.

Systems that don’t meet agency needs: Government’s focus on selecting the lowest-price bid that meets its requirements — within a fixed price — can sometimes result in solutions that aren’t quite right for the demands an agency will face in a shifting cloud-based environment. Whether it’s an application that doesn’t follow AWS best practices or an unnecessary storage system, working with a poorly planned solution can increase costs.

Lack of visibility: As cloud infrastructures grow, one of the biggest challenges is maintaining visibility into the enterprise. Agencies can lose track of abandoned cloud instances that still take up some of the budget. The complexity of the infrastructure also can make getting accurate information on spending difficult to come by.

Solution: Follow the Money

Cloud financial management is about knowing how money is being spent and what you’re getting in return by using effective metrics — which can be easier said than done. “Basically, they need to monitor the usage cost of the billing data,” Krishnaiah said. “You need to measure what you are managing.”

The scalability and the pace of agile development in the cloud, however, has given many enterprises the idea that the cloud isn’t really measurable. Krishnaiah disagrees. “You can measure” with the right tools and processes, he said. “You can deliver the project within a predictable budget.” Among the key steps involved:

Tagging: Implementing an effective cloud cost modeling and budget management system starts with assessing the system, Krishnaiah said. And that starts with tagging, using labels to apply customized metadata to cloud resources. Tags can be used on any type of resource — like user groups, databases or virtual machines — and make them easier to manage, search for and filter. They provide visibility into the enterprise and help optimize costs across cloud providers.

Controlling resource usage: Managing resource capacity is an essential element. Agencies need to be able to estimate the resources required for a cloud instance and shut them down if needed. Most government work is handled during business hours, but the cloud runs 24/7; plus, developers can spin out new capabilities at any time. Shutting down those resources for the night can save 1.5 times or more over the costs of business-hours consumption.

Monitoring: Monitoring, with data analytics and the dynamic ability to enforce resource limits, also helps keep usage and costs in line. Keeping track of billing by interactively and dynamically controlling resource usage is key to cost optimization.

This article is an excerpt from GovLoop’s report, “Rein in Runaway Costs Through Cloud Financial Management.” Download the full report here.