Service Management applications are rarely stand-alone systems. More often than not multiple integrations with other systems or data sources are in place. We have been building, hosting and supporting these integrations for our customers as long as we can remember.
Until recently we still did this the ‘old school’ way: for each customer we hosted a dedicated machine on which we installed the required software and code. Simple and effective, just the way we like it.
The world has changed, however. Hosting and maintaining servers is not our core business and there are more clever (and efficient) solutions for our requirements these days. So we decided to rebuild our integration infrastructure for the ground up.
We decided to go ‘full AWS’ and replace our hosted servers by Amazon Web Services components.
The image below provides a high level overview of the result;
“Who needs servers anyway?”
Our developers store the code for the integrations on the Github platform. This code is automatically deployed into Elastic Containers, fully maintained by Amazon. Scheduled tasks and scripts run on AWS Batch which is also maintained by Amazon. Each update to the code is immediately deployed to staging or production (depending on the code repository), providing a completely hands-free deployment for our developers. So we moved from hosted integrations to Continuous Integration and Infrastructure as Code, to just name a few buzzwords.
But the best news is that we save a lot of precious time and effort which we can use to keep up with the demand of our customers and can now fully focus on building rock solid while still affordable integrations.
