MIGRATE TO AWS
Our client Caple is the European specialist in SME growth credit. They provide this credit in close cooperation with BNP Paribas Asset Management. They interact with their SME clients through a network of independent and experienced professional advisors: their Partners. Caple supports them with technology and expertise in SME credit and provides the link to the institutional distribution capabilities of BNP Paribas Asset Management.
Since the beginning of 2020 Team Rockstars IT has been helping Caple in their IT challenges. As a trusted Business Partner, we strive to not only think and act in short term operational solutions, but always focus on value added advise and solutions for the long term, which contribute to our client’s business goals.
The challenges were among others primarily scalability and costs. The scalability problem was not so much due to peaks or extra load. It was a high-availability, auto-scaling and continuous delivery for production without downtime challenge! Costs were considerably due to the usage of a managed private cloud, new services/changes/maintenance for example networking, patches, back-ups en restore fell under the hosting provider and were billed.
Tech challenges: how to set up the “monolith” in such a modular way that we can grow to a microservices architecture in the future if needed. Another challenge is how we can migrate the production platform to another cloud without a lift-and-shift.
Migrate to AWS, via:
- Web hosting: usage of Route53 as DNS (Domain Name System) manager along with S3 (Simple Storage Service) and Cloudfront. This enables to host an Angular app;
- Backend: built a high-availability environment using ECS (Elastic Container Service), ECR (Elastic Container Registry), ELB (Elastic Load Balancing) and RDS (Relational Database Service). There was a choice to use Beanstalk or a lift-and-shift approach, but in the end we opted for ECS with load balancing. The reason for this was mainly the control we get with the ECS and ELB and cost savings. Beanstalk is easier compared to ECS, but it costs a bit more. In terms of cost savings, you can think of reserved instances for your production environment and spot instances for your test environment;
- CI / CD (Continuous Integration / Continuous Delivery): to implement CI / CD we chose github in combination with AWS code pipelines.
- High availability;
- Continuous delivery for production without down-time.
- Quickly boot up EC2 (Elastic Compute Cloud) instances to run prototypes;
- Security: Secrets, encryption data-at-rest, encryption data-in-transit, roles, policies;
- Using an AWS service without much effort;
- Unexpected business value: improved user experience feedback because of the stability and performance.
- Expected cost savings up to more than 50%.