Build vs Buy: Software Systems at Jurassic Park
summary
Jurassic Park builds custom systems (door‑lock IoT that fails in power outages, a three‑year implementation project that collapses) and discovers maintenance is the real cost — up to 90% of total lifecycle costs. A small security tool costs ~$9,000 to build vs $3,000/year to buy; a monitoring prototype (4 weeks to build) costs ~$7,600 plus ~$3,200/yr maintenance versus a $2,700/yr vendor. Vendor risk appears in the Raptor/event‑stream collapse; decisions hinge on vendor lock‑in, opportunity cost, and the Dilophosaurus factor.
AI-written summary · about
chapters (33)
- 0:00Intro
- 0:19Welcome to Jurassic Park
- 0:31An Uncomfortably Close Look
- 2:52Build is Unique
- 3:26Build is fun!
- 4:06Hire Ninja Developers
- 4:29Digging for the Details
- 5:31"Boot up the door locks!" Dr. Alan Grant
- 6:18Maintenance Time
- 7:06System Health
- 9:12Handling Outages
- 10:08Bus Factor
- 10:52Dilophosaurus Factor
- 16:15"We have all the problems of a major theme park and a major zoo" Ray Amold
- 18:00Maturity is overrated.
- 19:24Buying is Fast
- 20:31Buying Sucks
- 21:55Project Manager with a shotgun
- 22:39Vendors!
- 23:00Option 1: Industry-Standard
- 25:01Option 2: Highly Configurable
- 26:24Vendor Lock-In
- 27:14Sticky
- 28:40Not-Invented-Here Syndrome
- 33:33Mud on our Face
- 34:15What is the Impact of Failure?
- 36:25What is our bias?
- 37:18What could we do instead?
- 38:50Next Time!
- 39:55Hello Newman!
- 40:22JavaScript Error Monitoring an example
- 44:10Our Opportunity Cost
- 46:14Todd Gardner
related talks
30mBuild vs. Buy - a strategic decision for systems and people
28mPostmortem on the Ingen Incident
25mBuilding Applications in the Cloud
18mSecuring Open Source in Nomura’s Software Supply Chain
44mDemonolith the Monolith Think Again!
46mRugged Software Engineering
30mProduction › Prototype: How a Small Team Replaced a Core Platform in 90 Days
30mFueling The Future: Building Robust Engines