Platforms are brownfield

The new platform will fix all the problems of the legacy platform" are the famous last words of every platform engineer. Usually right after they’ve completed the first 90% of the migration and just before discovering the next 90% of the work yet to be done.

There are two primary sources for this failure mode. First is the misapplication of startup methodologies like lean software development that are effective in the market-based world of product development to the command economy inside large organization.

Second is that platform engineers relentlessly push to consolidate their platforms to simplify and remove incidental complexity. But this puts them at the risk of trying simplify the inherent complexity of the organization’s business domain, which will always resist being simplified.

Platform engineers can escape this fate by embracing brownfield development. In this talk, you’ll learn how to use existing shadow IT as a tool for market research. You’ll also learn a process for finding what “minimum” and “viable” really mean in a platform MVP by finding the “weird” teams early in the design process. You’ll hear stories of how exploring the reasoning behind zany deployments led to discovering hidden business requirements. And you’ll learn to replace the cynicism of resume-driven development with building trust with peer teams.



Tim Gross

Tim Gross develops open source infrastructure software. He is currently a staff engineer at HashiCorp working on Nomad. For the past two decades he’s been been a software engineer, sysadmin, ...