How Houghton Mifflin Harcourt Went from Months to Minutes with Infrastructure Delivery
summary
Claim: infrastructure delivery shrinks from months to minutes by building a Mesos-centered self-service platform. Implements Apache Mesos + Apache Aurora, containers, Salt for orchestration and Terraform for infra-as-code. A seven-person ops team supports ~350 engineers running 1200 applications with automation, bedrock shared services, and continuous delivery instead of 6–18 month releases, saving six-figure annual costs. Emphasizes people work: embedding with teams, enforcing expectations via inspection, post-mortems, and a recent Aurora incident that revealed an 8-hour job-db gap driving Terraform-backed recovery plans.
AI-written summary · about
chapters (4)
related talks
42mHow Houghton Mifflin Harcourt Went from Months to Minutes for Infrastructure Delivery
25mFrom Hours To Minutes: The Evolution of Platform Engineering at Decathlon
42mUsing Mesos to Drive Devops Adoption at Scale
42mA Year with Apache Aurora
31mContinuous Delivery Lessons Learned Through Hypergrowth
27mAdvanced Multipurpose Infrastructure: A Self-Managed Cloud Environment with HashiCorp Terraform
31mDevOps in Infrastructure — When It's Mostly Ops...
29mWhat would you do with 8k extra hours? Turning infrastructure into a strategic advantage