ContainerizeThis 2016

Have a downloadable infrastructure software product, but an audience which is accustomed to instantly-available hosted services?  Why not make an online trial to lower the barrier to adoption?

This is the question we explored for our database product, Neo4j.  Plenty of questions came up - how do we authenticate users, isolate them, and scale the service while providing a genuine product experience? What types of data can we provide for users that is both meaningful and interesting?

Docker and AWS came to the rescue -- enabling us to launch isolated containers for each individual user in a time-limited trial. 

This session will discuss the architecture of that implementation (queuing services, python, elastic container service, data ingestion) + lessons learned during the implementation (and first week in production!).  We'll review the strengths and weaknesses of our approach along with later explorations of using Docker Cloud for a similar purpose.