Every new approach needs a catchy name. During this session Dave Hall will introduce the concept of "environmentless workflows". Serverless architectures still have servers behind the scenes. In environmentless workflows there are still environments in the workflow, but users don't have to think about the environments anymore.
Today teams need to iterate quickly. Releasing once at the end of each sprint can be too slow. Teams should be able to release when work is ready to go. Some times due to unavoidable dependencies, multiple changes have to be packaged up and deployed together. If one of the pieces has been held up or fails QA it can block everything. Instead of waiting it should be possible to drop the piece that isn't ready, retest and deploy.
Environmentless Workflows do away with the traditional idea of fixed dev stage and prod environments. Instead all work starts in sandbox, which is an isolated clone of production with an associated feature branch. Once work is complete the environment can be cloned to perform automated testing. If the tests pass the changes are available in the pending work list. All pending items are able to be added to builds. The builds are a combination of one of more changeset. These are then tested, if the tests pass the environment is made available to acceptance testing. If the build is approved it is then is promoted as the new production environment. This session will be 10% awesome slides, 10% history, 20% theory/overview, 50% walk through, 5% failing demos, 10% Q&A, 2% math and 1000% fun.
By the end of the sessions attendees should have a good understanding of what Environmentless Workflows are, what's needed to implement and the benefits for their teams. The session assumes that participants will have an understanding of continuous delivery and docker.
Note: I am still building out the services to support the workflow discussed in this talk. It will be incomplete when presented at this event.