If you're even remotely interested in the convergence / VoIP / unified communications field and haven't read Steve Taylor and Jim Metzler's article on SIP interoperability, you should. I would like to expand a bit on why SIP is so incredibly important, and why the future of IP communications depends on it.
In my honest opinion, we have a quiet, yet major problem developing in the industry. There are simply too many systems and environments that are utilizing proprietary or "modified" protocols for signaling. So, development in this area has been vendor-centric and closed, and not towards a common and unified technology.
Take a close look at the technologies in the user-facing telecommunications field that have been extremely successful. In a recent blog post, I discussed the past, present, and future of the PRI. What is hidden behind the success of this trunking technology is something that is key: all major telecommunications platforms can interface to, and "speak", PRI. This example proves one thing: if we're going to promote an interoperable industry, then we need to be, well, interoperable.
SIP is a great thing. Vendors who take the core protocol set and add their own proprietary enhancements or deductions are endangering the future of the protocol. The standardization process of SIP is too slow, simply put. Even today, the industry seems to move backwards in SIP development, detracting from the common goal of being an "interoperable" protocol. The open-source-focused platforms (Asterisk, sipX, Sip Express Router) are doing the right thing, even if it requires additional development in "assisting protocols" for feature enhancements, etc.
Until we can get the "big boy" vendors on board, we're stuck. If we're going to have a "standards-based-protocol", then let's keep it inter operable. With vendor proprietary modifications, is it really SIP?
The Leader in Network Knowledge
SIP is the new H.323
Seriously, can you cite an example of any big bad vendors violating the SIP protocol? It is pretty loosely written, so there are many opportunities to not interoperate without anyone violatin', just due to differences in interpretation. Most of the issues have now moved "up the stack" to supplementary features like hold, transfer, conference, park, etc. It would be great if everything worked like we want, but get over it, life isn't like that.
Apparently nothing was learned from H.323:
1) Very poor interoperability beyond making a phone call (even that doesn't always work between vendors)
2) Actively firewall hostile protocol at the transport layer, complicating network design.
3) The ability to claim compatibility or conformance on a feature list while interoperability is pathetic.
It's almost like the vendors want to sell their proprietary phones, so they have no incentive to make a cheap generic SIP phone work well with their call control. Just where do you think the bulk of the revenue comes from, the phones or the call processing servers?
When customers start to vote with the $$s, the vendors will change. The real question is why people are still installing PBX style systems in the age of the cell phone. Ooops, might put myself out of a good paying gig....