FutureLearn: Boosting Business Value with Cloud Cost Optimisation
Published: February 2024
Client Name: FutureLearn
Industry: Online education and educational technology (EdTech) provider.
Executive Summary
The main challenges for the client were increased cost and complexity of its AWS infrastructure that grew over time with the expansion of the business. Irysan supported FutureLearn in investigating the root cause analysis of the increased cost, specifically looking at the resource utilisation and spending patterns. We discovered that AWS costs were escalating due to several factors, which included suboptimal resource allocation, lack of cost visibility and reactive cost management practices.
Client Background
Launched in 2012, FutureLearn has become a leading online learning platform, democratising access to education through high-quality courses from top universities, such as King’s College London and University of Edinburgh, and organisations worldwide. Since launch, the company has grown to over 20 million registered users, a catalogue of over 2,000 courses, and partnerships with over 260 institutions.
Recently acquired by Global University Systems, FutureLearn continues to expand their offerings and innovate in online learning formats.
FutureLearn relies heavily on cloud infrastructure provided by Amazon Web Services (AWS) to power its product offering
Challenges
The main challenges for FutureLearn were the constantly increasing complexity and costs of their AWS cloud infrastructure. The problem was further exaggerated by the lack of cost visibility and reactive cost management practices. FutureLearn didn’t have a clear visibility of its AWS spend, which made it difficult to identify and address cost-saving opportunities. Furthermore, the company utilised on-demand pricing combined with absence of practice of resource cost monitoring.
Solutions
During the project, Irysan worked with FutureLearn to identify both direct savings initiatives and best practice optimisations which, while not immediately reducing the AWS bill, are essential for sustainable cost control.
Irysan has found that the costs of the AWS pre-production environment, CI and monitoring have been increasing year-on-year, as a result of new developments. We suggested tagging resources for an improved understanding of the cost associated with each project.
Additionally, we found a large volume of old snapshots, which were promptly deleted. FutureLearn had a practice of keeping too many daily database snapshots with no real use case - we suggested reducing the number of snapshots and the duration of time for which they are kept.
Finally, the Irysan team found that the main production database was oversized, and could be downsized by 50% without impacting the business.
We investigated switching EC2 clusters to Fargate for increased efficiency and reduced complexity. This also allowed us to implement a granular backup strategy.
Once the key optimisations have been implemented, we suggested further lowering costs by reserving AWS capacity and utilising the AWS Enterprise Discount Program.
The outcomes achieved
The efficiency-focused interventions collectively contributed to a substantial reduction in the cost and complexity of FutureLearn’s AWS infrastructure, reaching system optimisations of up to 50%, cost efficiencies of 32% and future cost savings associated with reserved pricing, which could now be applied to a lean infrastructure architecture.
The simpler AWS infrastructure landscape is now more maintainable and requires lower maintenance overheads. Importantly, with the new practices of resource tagging and transparent spend with Application Manager in place, FutureLearn team is now well-equipped to monitor and address future increases in complexity and cost.
Results and Benefits
When working on this project, the chosen approach of “going back to basics” and employing discipline in understanding the client’s AWS resource allocation patterns and consumption needs served our client very well.
Tagging resources used by software developers brought clarity and transparency of development costs in AWS. Snapshot monitoring and removal of a large volume of old snapshots further contributed to cost reduction and an overall leaner infrastructure. Reduction of the size of the production database had a similar positive impact.
One important result of this project was that development and the SRE teams felt more in control of their resource utilisation and real resource needs.
Conclusion
Increases in technological complexity and spend are a normal occurrence for evolving businesses. Even when SRE teams adhere to best engineering practices when designing and implementing their cloud infrastructures, cost monitoring is often considered only when it becomes unusually high.
In this instance, going “back to basics” and methodically reviewing which resources are necessary and which are obsolete, and evaluating the actual vs provisioned resource allocation was the right approach to firstly understanding, and then removing unnecessary resources and downsizing over-provisioned instances.
The financial savings resulting from this project amounted to £ thousands per month, which demonstrates the significance of the impact that cloud cost management can have on a company's financial health and growth potential.
