As a front-runner in cloud technology our goal is to always build low maintenance & low-cost digital platforms to speed up the time to market for our clients. This blog article describes as to how we provided one of our DIY clients with a modern cloud e-commerce architecture, that contributed to reaching their business’ KPI’s.
Rolling out new features for the huge e-commerce platforms we maintain used to be a struggle, mainly due to complex and heavy software stacks making use of SAP Hybris. To make a platform more agile we kicked-off with an AWS cloud migration about two years ago. This lift & shift migration lowered costs and gave us more flexibility from an operations point of view. Nowadays, for example, temporarily adding new environments for testing purposes just takes no more than a few minutes.
Starting with a lift & shift migration was just the first step in our cloud journey with our client and all involved partners. Optimizing for the cloud was the next big thing. Our client wants to be able to roll-out small changes without big-bang releases. Our joint approach of choice was based on a micro-services strategy. The idea behind this strategy is to isolate different key functions, host them on separate virtual machines and use API’s to fulfill the communication between the services. With this foundation layer, we were able to make changes to different functionalities without ever touching other business critical software components. At this point different micro-services had been built for the checkout process, price calculation, stock services and image processing. As a result, the software stack transformed to a total of 10 different micro-services, which gives our client a huge advantage in rolling-out new features fast and free of risk.
Interested in the more technical details? Our stack contains Atlassian Bamboo for CI/CD, Rundeck for scheduled jobs and management of clusters and services. For monitoring insights we rely on AWS X-Ray, New Relic and CheckMK. Under the hood, you will find a lot of Beanstalks with a slight taste of Docker, powered by Redis, AWS DynamoDB, AWS DynamoDB DAX, AWS CodeDeploy and Netflix/eureka, to have it all work together in harmony. All is created using AWS Cloudformation-based templates.
The transformation from a classic hosting infrastructure towards a cloud native micro-services based platform for one of our DIY clients resulted in:
- Being able to__speed-up-the time to market__ by a fine-grained deployment process thanks to a micro-service approach;
- No downtime. A very stable platform which achieves 100% uptime performance;
- Auto-scaling. Again, thanks to different services, scaling takes places at different functional levels;
- A__performance increase of 20%__ from March till June, compared with the same period in 2016.