How we use Technology from AWS to Provide Business Value For ISL.

18/11/2019

Here at Cucumber we have been working with ISL for many years, helping them along their journey to digitisation and cloud technology. For more information about their journey and business needs, please see this blog post. This article will talk about the technology we have used in building their application, and how the technology from AWS has greatly contributed to the success of the project. 

The initial application we built for ISL was a cloud native application, hosted on one of the leading cloud vendors. However, it immediately suffered from performance problems, primarily due to high peak loads on the database (and artificial IOPS limitations by the cloud vendor). Luckily these high loads were known in advance, and so we were manually handling the situation by scaling the database performance to handle the high loads (and then scaling back down to save cost). Obviously, this wasn't a sustainable solution, so we began to evaluate different technology and hosting solutions. AWS was a clear leader, we particularly liked their embracing of open source technology, and their range of PaaS offerings with no artificial limitations or throttling.  

The entire application was moved to AWS, and switched to using an open source stack of Linux servers and RDS PostgreSQL for the database. This transition was surprisingly straightforward and included containerising the application in Docker. Almost overnight the performance problems disappeared, and the overall hosting costs were lowered (as there were no more commercial licenses). So from a purely financial and performance viewpoint, the transition to AWS was a great success. 

Since then the application's scope has expanded, and we have turned to AWS services to provide the required functionality. This includes things like S3 for storing images (now numbering in the hundreds of thousands), Lambda to provide stateless de-coupled functionality (such as image resizing) and QuickSight for Business Intelligence (BI) reporting. By utilising these existing AWS services, instead of building them ourselves, we could rapidly bring new features to the application to meet market demands. 

More recently, the entire application has been moved into a Kubernetes cluster, provided by AWS EKS. Having previously built and managed a Kubernetes cluster manually, the move to EKS removes a large maintenance and management overhead. The database has been moved to RDS AuroraDB, which provides even greater performance.  

While ISL was our first application to be hosted in AWS, we now have several customer applications in our AWS EKS cluster. AWS provides us with cost effective, performant and reliable services, and allows us to focus on our primary goal of delivering value to the customer. 

Written by Guest Blogger

Previous
Previous

Approved R&D Research Provider tax status

Next
Next

Exploring environmental innovation at Matariki X