Untold story
5 likes646 views
The talk discusses the history of Vyatta, an open source networking company founded in 2005. It aimed to be the RedHat of networking using a free software model. However, Vyatta struggled with issues like lack of infrastructure support, scaling challenges, dependence on outdated routing protocols, cost of support, and competition. While the free software model provided benefits like free developers and publicity, Vyatta failed to gain significant market share. The talk suggests free software platforms take time to mature and that new opportunities may arise from technologies like DPDK, Yang models, Openstack and OpenDaylight.
Untold story
- 1. The untold story of free software in networking Stephen Hemminger @networkplumber
- 2. Vyatta's Journey ● History ● Plan ● Potholes ● Lessons learned
- 3. History ● Founded 2005 ● 2007 Debian platform ● 2008 XORP Quagga→ ● 2012 Brocade acquires Vyatta ● 2013 Community version ends ● 2014 vRouter 5600 (DPDK) released
- 4. “Plans are of little importance, but planning is essential” -- Winston Churchill
- 5. Free Software + Free Publicity + Free Developers ---------------------------- = Success
- 6. 100 75 50 25 0 Marketshare% Innovators 2.5 % Early Adopters 13.5 % Early Majority 34 % Late Majority 34 % Laggards 16 % Diffusion of Innovations
- 7. Free Software gets attentionFree Software gets attention
- 10. Free Developers ● Worldwide users ● Bugzilla ● Forums & Chat ● Easy to build – Git – ISO
- 11. Best Documentation was from the community!
- 12. Developer dream
- 14. Missing Infrastructure ● No Commuity Manager ● No Foundation ● No Contributor Agreement ● No Pull Request process
- 16. Roadblocks ● Scale ● Effort ● Features
- 17. Routing at scale
- 18. Dataplane Performance 64 128 512 1518 0 1000000 2000000 3000000 4000000 5000000 6000000 Linux DPDK Packet Size FramesperSecond Line Rate
- 19. Dependence on extinct Quagga ● Missing OSPFv3 ● BGP threading issues ● BGP attributes ● Not enough love
- 20. http://headrush.typepad.com/creating_passionate_users/2005/06/featuritis_vs_t.html
- 21. Where are the developers?
- 22. Final knockout blows ● Real cost of support ● Double the release effort ● Per unit licensing ● Self competition ● Two routing protocol stacks
- 23. “Success consists of going from failure to failure without loss of enthusiasm” -- Winston Churchill
- 24. Vyatta was in the valley Time Legacy Hardware Software Networking
- 25. Open Ecosystem is awesome ● Tools – git, perl, python, ... ● Infrastructure – Debian, Linux ● Surfeit of tools
- 26. Package vs Platform ● Packages – Multi-vendor – Decoupled ● Infrastructure – Process – Roadmap – MAINTAINER?
- 27. New Technolgy = New Oppurtunities ● DPDK ● Yang ● Openstack ● Open Daylight
- 29. Free Software maturity + Software networking hype + Multi vendor collaboration ---------------------------- = New IP
- 30. Questions?
- 32. The untold story of free software in networking Stephen Hemminger @networkplumber When you start on a trip, you end up in a totally unexpected place. Sometimes it's the journey that teaches you a lot about your destination. Read more at http://www.brainyquote.com/quotes/quotes/d/drake55 1817.html#XYhSypwOg9T65HRT.99
- 33. Vyatta's Journey ● History ● Plan ● Potholes ● Lessons learned This talk covers the history of Vyatta, what happened, what worked and what didn't
- 34. History ● Founded 2005 ● 2007 Debian platform ● 2008 XORP Quagga→ ● 2012 Brocade acquires Vyatta ● 2013 Community version ends ● 2014 vRouter 5600 (DPDK) released
- 35. “Plans are of little importance, but planning is essential” -- Winston Churchill
- 36. Free Software + Free Publicity + Free Developers ---------------------------- = Success Vyatta wanted to be the RedHat of networking. Networking is dominated by mature proprietary legacy vendors (mostly Cisco)
- 37. 100 75 50 25 0 Marketshare% Innovators 2.5 % Early Adopters 13.5 % Early Majority 34 % Late Majority 34 % Laggards 16 % Diffusion of Innovations
- 38. Free Software gets attentionFree Software gets attention The fundamental issue when breaking into established market is getting noticed.
- 41. Free Developers ● Worldwide users ● Bugzilla ● Forums & Chat ● Easy to build – Git – ISO
- 42. Best Documentation was from the community!
- 43. Developer dream
- 45. Missing Infrastructure ● No Commuity Manager ● No Foundation ● No Contributor Agreement ● No Pull Request process
- 47. Roadblocks ● Scale ● Effort ● Features
- 48. Routing at scale
- 49. Dataplane Performance 64 128 512 1518 0 1000000 2000000 3000000 4000000 5000000 6000000 Linux DPDK Packet Size FramesperSecond Line Rate
- 50. Dependence on extinct Quagga ● Missing OSPFv3 ● BGP threading issues ● BGP attributes ● Not enough love
- 51. http://headrush.typepad.com/creating_passionate_users/2005/06/featuritis_vs_t.html
- 52. Where are the developers?
- 53. Final knockout blows ● Real cost of support ● Double the release effort ● Per unit licensing ● Self competition ● Two routing protocol stacks
- 54. “Success consists of going from failure to failure without loss of enthusiasm” -- Winston Churchill
- 55. Vyatta was in the valley Time Legacy Hardware Software Networking
- 56. Open Ecosystem is awesome ● Tools – git, perl, python, ... ● Infrastructure – Debian, Linux ● Surfeit of tools ● Free software platforms like Linux take time ● Free software costs can be spread across multiple vendors ● When needs change, users change
- 57. Package vs Platform ● Packages – Multi-vendor – Decoupled ● Infrastructure – Process – Roadmap – MAINTAINER?
- 58. New Technolgy = New Oppurtunities ● DPDK ● Yang ● Openstack ● Open Daylight
- 60. Free Software maturity + Software networking hype + Multi vendor collaboration ---------------------------- = New IP
- 61. Questions?