Change for the Better

From limited deploys to hundreds per week

Four years ago, ITHAKA had large development and QA teams; quarterly releases; long planning cycles; one monolithic application with resource intensive maintenance needs; limited accountability for the operationalization of said application; data maintenance tasks that took weeks to complete; and a small on-call team. Our more traditional IT processes constrained our ability to deliver as fast as we wanted to — small changes could take months and large projects could take a year to get to production.

Today, ITHAKA runs many, small, multi-disciplinary, dedicated engineering teams; 200+ applications in production; an on-call schedule with 70+ people responsible for their own applications; a pipeline handling deployments to production many times per day with minimal planning and no centralized control. Product owners can identify ideas for new product features in the morning and measure results with real users in the afternoon. Our private PaaS on AWS enables our engineering teams to operate independently and get things done.

How did we get from where we were to where we are? We wrote a lot of documents and talked to a lot of vendors but our most important work was to demonstrate key results — for example, using AWS to complete work in an hour that would have taken weeks on the old platform. Early successes earned us additional resources to do more — we built a web interface for application deployment that enabled any engineer to deploy a new app as fast as AWS could launch instances. Our continued successes — and learnings from failures — enabled us to grow.

We’ve learned an incredible amount during this transition, and continue to learn every day. Many facets of that learning deserve their own conversations. In time, we’ll talk about how we’re evolving both technically and culturally to create our new reality.