Last week we carried our thoughts and impressions from the VON (voice over networks) conference in London and focused on the difference between the instantaneous gratification of Skype for private individuals and the safe and efficient enterprise wide VoIP implementations, reliant mostly on the SIP protocol.
Rohan Mahy, co-chair IETF SIP and SIPPING Working Groups, took issue with some of the things that Niklas Zennstrom of Skype and Kazaa fame said at the conference, and although enterprise VoIP isn't a central technology to Faultline, it is a valid and radical cost pressure on telcos and an opportunity for various vendors.
Much of the irritation that Mahy showed at the VON conference relates to the Zennstrom view that SIP is a poor protocol because of its inability to handle firewall and NAT traversal. So in Mahy's own words we thought we'd offer the other side of the argument.
"There are very few people in the VoIP industry who understand firewall and NAT traversal well. Lots of people in the VoIP industry think that they have some technique which is new and good. Unfortunately the parts that are new are rarely good and the parts that are good are usually not new. Explaining to these people what is technically broken with their proposals and why is an extremely time consuming process, and is repeated every time a handful of new companies start to go into the operational phase. As a result, the folks who understand these issues either go on a holy war each time a 'new great NAT solution' crops or just ignore the new folks.
"Implementers want a one-size fits-all solution where one does not exist. I can describe significant limitations with every approach, but by implementing a handful of these tools and applying the best tool for the situation, you can do much better than just implementing one. When an implementer comes to me or Jonathan Rosenberg, or Jon Peterson and asks how to do NAT or Firewall traversal, most of them are frankly not willing to invest the attention and energy needed to implement real solutions to this problem for VoIP (not for SIP, but for VoIP).
"I believe that Niklas [Zennstrom of Skype] is in this category. He is not willing to admit that his solution is completely broken by firewalls that try to block p2p music sharing, that the algorithms which gives his product good sound quality can be implemented on any product, and that the supernode model not so occasionally results in very bizarre routing (example: a call within Europe going through Australia)
"As with anybody who has technical problems with IETF protocols, I invite them to write a description of what problem they are trying to solve which they feel is not addressed by existing protocols and provide some motivation. Zennstrom has not provided any rigorous analysis or even detailed explanation of these alleged technical problems.
"The traditional telecom folks want to reproduce an environment which is familiar to them. SIP does things differently, but many folks generally stop there and don't try very hard to understand how to accomplish their goals 'the-SIP-way'.
"I believe that Zennstrom has a different motivation. He is providing a packaged service and I believe that he is afraid of the idea of open services and open protocols, because these things directly threaten his business model."