Things to Consider When Optimizing Cloud Infrastructure Costs

Why is cloud cost optimization crucial?

As your organization moves more and more workloads to the cloud, it’s important to understand the impact of successfully and cost-efficiently running your applications in a cloud environment (or even across multiple clouds). Optimizing your cloud deployment isn’t a ‘one-time-thing’ that you can implement in a day or a week and be done with it. Strategies for optimizing your cloud infrastructure costs can range from those that are tactical and easily executed in a matter of days, all the way to those that are more strategic and continuous in nature and ultimately support the organization’s bottom line. Streamlining your cloud spend has benefits that go beyond cost reduction. By guiding your technology and infrastructure decisions, it can positively impact developer productivity and help future-proof your tech stack.

Especially in times of economic uncertainty, development teams and technology organizations are often requested to provide the same level of uptime, security and resiliency for their services while reducing the overall TCO (Total Cost of Ownership). So cloud cost optimization is an increasingly important topic.

No matter if you’re using the Aiven Cloud Data Platform for the first time or scaling your existing deployments, the following sections will help you. We share best practice for ensuring optimal usage of your cloud resources and a framework to help you ask the right questions and make the appropriate decisions for your overall architecture.

Cloud cost optimization strategies: immediate steps

When you are looking to streamline your cloud spend, there are a few easy steps you can take from the start. Some of these include:

1. Verify that your services are properly sized for the volumes of data you are expecting. Understandably, development teams often want to overprovision some of the services to accommodate spikes in data traffic and mitigate the possibility of downtime. However, correct sizing can have a significant impact on the cost structure. Evaluate whether any unnecessary cloud resources are being utilized and, where appropriate, downsize those to the correct level or terminate any unused resources altogether.

2. Try to minimize any aging, homegrown configuration code and pieces of your infrastructure that are hard to maintain. More often than not, these configuration elements are prone to mistakes and bugs and need both additional governance and extra time to provision and tear down. This results in lower developer productivity through additional man hours and unnecessary work for your developers. The solution? Ensuring that such configurations can be automated and embedded into your tooling. This can be done by using a Terraform provider or making sure that you manage any Kubernetes clusters and resources with the use of a Kubernetes operator for maximum automation and efficiency.

3. Finally, ensure that the appropriate monitoring and alerting configurations and best practices are in place. This will ensure you can easily detect abnormalities in your resource consumption and provide a real-time view of your cloud resources consumption at any given point in time. Bringing the correct monitoring and metrics components to your tech stack is a must-have when you are streamlining your cloud spend.

Cloud cost optimization strategies: a midterm approach

Now it’s time to step back and take a look at the bigger picture of your overall architecture and tech stack. This is a midterm implementation strategy where you need to find answers to the following topics:

1. What do your operational patterns look like now, and can they be optimized further? For example, are there spikes in utilization that you can predict based on specific company activities and plans in the coming months or quarters? Do you experience any seasonality in your sales or operations? Answering such questions will ultimately help you define an optimal strategy for provisioning your cloud operations for increased efficiency

2. What do your dev, Q&A and production environments look like right now? And can they be further streamlined? Ensuring that different types of instances are treated appropriately in your cloud infrastructure will make sure that your cloud resource pool is used optimally. Here you should consider using automation and IaaC tools such as Terraform, AWS CloudFormation, Azure Resource Manager, Google Cloud Deployment Manager and others.

3. Which components of your overall tech stack need to be evaluated? Are there proprietary, commercial pieces in your infrastructure that can be easily migrated to open source alternatives at a fraction of the cost? What is the cost of operating and maintaining these proprietary pieces of infrastructure and how can this be reduced by turning to open source alternatives (especially when they are fully managed for you out of the box)?

   Be sure to thoroughly examine and evaluate the different components of your architecture. Task your teams with finding alternative solutions that can do the same job possibly at a more competitive price point.

   Additional benefits for choosing open source alternatives for your tech stack include getting access to a pool of developer talent from the community, having the opportunity to reuse some of your data assets and improving your developer productivity with familiar development tools used and developed by open source communities.

One company that has successfully implemented this is ‘clean’ digital asset data, insights, and research provider Digital Asset Research. The team at Digital Asset Research reviewed specific components of its architecture and by moving to Aiven they were able to secure a 99.99% uptime, lower their costs by 25% and improve their performance by 10x when scaling their operations.

Cloud cost optimization strategies: an ongoing team effort

Now it’s time to look more strategically into what your tech stack will look like for a sustainable growing operation in the years to come. Here you should take into consideration different components such as:

1. Your cloud vendor. Which cloud vendor are you currently using and how happy are you with the level of service and cost provided by them? How will your cost structure be impacted once your services and applications expand to larger scale? Here you can evaluate whether moving to a different cloud vendor would bring reduced costs or improved service levels if your deployments increase in the future. You should also factor in other cloud migration costs — what downtime will you experience and how will this migration process impact your overall team productivity in the midterm?

   Aiven supports a painless migration process to a different cloud vendor or region with minimum to near-zero downtime. Additionally, Aiven has a global presence with 100+ supported cloud regions and can support you wherever you need to operate. An additional strategy could be staying with the same cloud vendor but choosing a different managed service to reduce costs and improve your overall cost predictability with a vendor like Aiven.

2. Your technology stack. What tech are you currently using for your different use cases? And is it optimal? Is there a different piece of technology that can both lower your data management cost structure and better fit your specific use case(s), for example in terms of latency requirements or networking traffic volumes you expect in six or nine months from now? Thoroughly examine your architecture and ensure that the chosen tech stack is future-proof and can support your organization in the quarters/years to come.

   Aiven’s easy migration process allows you to move to a different, open source managed service that can support your growth without exploding your cloud expenditure. Aiven services are deployable in the cloud of your choice, so there’s no need to change or move away from your hyperscale cloud provider if you choose to leverage any multiyear spend commitments you have secured with them already.

   A company that has done it successfully is delivery service Wolt. Wolt has been investing in its technology stack, building a platform that can scale for quick expansion and future growth. “One of the key enablers for us to compete successfully in the delivery space is the fact that Wolt has been investing heavily in the platform and the technology,” said Mikko Peltola, Head of Infrastructure at Wolt. “We’ve invested in building a platform that supports quick expansion and growth for our business.”

3. Last but not least your team. Does your current code and technology limit your developer productivity? How can you make your developers both more productive and successful with the chosen suite of technologies they are using? Are they happy operating and maintaining pieces of software that are aging fast? Or would they rather deploy open source technologies backed by vibrant and live communities that they can become a part of and contribute to? Ensuring that your tech stack meets (or even exceeds) the needs and demands of your tech team will significantly improve your team’s productivity and reduce your talent turnover, giving you a competitive advantage in the challenging recruitment market.

Wrapping-up

Optimizing cloud infrastructure costs is often seen as a quick fix rather than a strategic, long term and ongoing effort. While there are immediate steps you can take to reduce costs quickly, the mid to long term strategies are critical to ensuring a sustainable and cost-efficient mode of operating as a successful technology group.

Our team of experts at Aiven runs regular workshops with customers on what a potential cost optimization strategy could look like — tailored to the use case, existing tech stack and setup. If you are interested in finding out more about how your services and cloud spend can be further streamlined, get in touch with our team today and schedule a call.

Not an Aiven customer yet? You can schedule a demo and explore one (or multiple) of the 11 open source technologies provided in the Aiven console. Be sure to check them out.

Further Reading

• What is ETL?