Skype Journal Test

Saturday, September 20, 2008

A few thoughts on Skype interop tradeoffs

Interoperability comes at the expense of innovation. Once you have multi-network interop, you can't drag a whole industry with you as you improve technology.

For example, Skype keeps improving their codecs with a team of PhDs in Stockholm. The team (still hiring) updates Skype's proprietary codecs, and their configurations, to the whole Skype network every few months. Same with Skype's p2p engine configuration, and with NAT traversal as new routers and firewalls are discovered. So the technology that makes Skype work is changing often. As are the computers and networks Skype runs on. Skype learned things at 100 million user accounts it didn't know at 1 or 10 million. What will they learn by 500 million users? Time and change are giving Skype a unique depth of experience, knowledge and skill.

The move to mobile complicates this further. What Skype knew about networks, latency, UI frameworks, etc. all came in the PC/Internet context. Codecs and encryption that work easily on PCs will melt most phones. Wideband audio, stereo, spatialized audio? These crowd pleasers wait for several cycles of Moore's Law and years of mobile device hardware evolution. High def video? A pipe dream for the next decade.

Yet Skype must blend mobile usage into their PC network while keeping core values and brand notes. And Skype hasn't started to plug-in to web sites,  another front with its own changes and revolutions. 

So the technology is fluid, complex, in motion. But Michael Robertson's call for Skype to interop is a call to stop innovation and adaptation. Kids born when the conventions MR wants Skype to support were defined are now starring on Are You Smarter Than A Fifth Grader? This is like locking in mileage and pollution standards from the Mad Men era instead of reaching for the zero emissions 100mpg car.

So:

  1. At what point should Skype give up competitive advantage for the increased network effect?
  2. When the environment continues to change rapidly, when should you stop innovating and start commoditizing?
  3. Can you interop on some features but not on others? Which ones? And when is it worth it?

tags: , , ,

Follow Phil Wolff on Twitter or FriendFeed or on Skype.

Labels: , , , ,

0 Comments:

Post a Comment

Subscribe to Post Comments [Atom]

<< Home