Keynote: Make SDN Real
summary
Built a working stack: an OpenFlow-rooted controller (NOX controller), a distributed state platform (Onyx distributed state platform), and a virtual networking product that becomes NSX which reaches $1B run rate. Adoption concentrates on three use cases roughly 40% automation, 40% security, 10–15% other, hardware features prove hard to change because refresh cycles are 5–7 years, and key lessons are product before platform, virtualize before changing abstractions, and prioritize software and go-to-market over reinventing infrastructure.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 0:46Background
- 1:20Original Pitch
- 2:25Initial Idea
- 3:00Knox
- 3:32Networking is not computing
- 5:40Graph
- 6:20Value
- 8:09Virtualization
- 8:49Insertion
- 9:21Hardware
- 10:21Operations Problem
- 10:58Network Complexity
- 13:17Slicing
- 13:57NSX
- 14:52What did we achieve
- 15:10High level goals
- 15:40Things that in retrospect may be obvious
- 16:37Minimum time to get something done
- 17:13Takeaway
- 17:33Key Lessons
- 18:17Success is over determined
- 19:51Balancing two types of human nature
- 20:56Personal question
related talks
31mKeynote Panel: SDN@10
13mKeynote: Open Networking 2.0 – Raising the Game for the Networking Industry
26mRedefining the Networking Industry Through Collaboration
29mOpening Remarks & The State of the Industry
35mKeynote Panel: SDN/NFV Deployment Progress - Ecosystem Perspective
20mKeynote: Driving Network Automation and Cloud Native
23mKeynote: Bring Disruptive Innovation to Deployment with Curated Open Source
22mKeynote: Accelerating the Pace of Innovation