Daily Payload
Home | News Archive | Search | Submit Story | Subscribe | Contact Us

SIP Sucks. It Really Sucks.

May 17, 2009

I thought that headline might get your attention. No, we've not changed our minds about the incredible potential for IP Multimedia Communications. Nonetheless, this headline echoes the sentiment of so many people in the VoIP industry who have worked with the communication protocols that surfaced as we made the transition from PSTN systems to IP-based systems. In fact, the headline is a direct quote from an esteemed associate in the business who said, "SIP sucks. It really sucks. But, it's what we have."

Back in the early days of VoIP when SIP was initially published as a draft document (1996) in the IETF, there was a lot of excitement about the protocol and how it would revolutionize the communications industry. Well, there is no question that VoIP has revolutionized the industry, but not exactly the way some of us had expected and it certainly was not SIP that led the charge, either — H.323 led the market, took a very large share of the market, and still dominates in many spaces like videoconferencing. What SIP did was create new opportunities for those who wanted to get into the voice business, while slamming the doors on some who had long been in the business. What SIP was supposed to have done was significantly change the way in which we all communicate. To that end, it has not been a phenomenal success.

As evidence, what is the single most popular VoIP product in the world? Without question, it is Skype: a proprietary VoIP client with an active user base of about 17,000,000 users each day, and with well over 220,000,000 users registered. It offers one the ability to place voice and video calls, as well as the ability to send text messages and files. Without any hesitation, we can say that Skype has been quite a success and, most importantly, "it just works."

We cannot say the same with such enthusiasm about SIP. Now more than 13 years since its introduction, what great new capabilities do most SIP users have? Sadly, they just have voice service just like they had in the 1980s. In short, very little has changed. There are a few who are using video, but the number of such users is so incredibly small at this point that we cannot declare SIP as a success in the video communications space, either. Perhaps more importantly, where SIP is used for video communication, the business model is similar to that of legacy videoconferencing equipment of the 1990s. That is not necessarily a bad thing: IP provides considerably more bandwidth and video quality is, therefore, significantly better. Perhaps with the passing of time and continued increase in bandwidth, SIP-based video conferencing will be so ubiquitous that it will be like the voice telephone was of the 1980s.

Still, SIP sucks and it is not because it provides us with yet-another-telephone or yet-another-video-system, but the fact that it has miserably failed to significantly change the way we communicate. SIP allows vendors to create IP-based phones, but what about application sharing, transferring files, electronic whiteboards, or other kinds of applications that really take advantage of the IP network? How about allowing developers build new kinds of applications that one can use when communicating with his SIP device? The problem is that SIP simply was not designed for that: what you get from your SIP equipment vendor is what you get and all you'll get. SIP has a complex monolithic protocol stack design that is not open, much like the legacy PSTN protocols we tried to move away from. Further, carriers are deploying network infrastructure that is at least as complex as the legacy PSTN that existed before it. And for what? To allow us to make a phone call?

While SIP was heralded for years as a technology that would enable all kinds of new applications and allow users to do things that were never possible before, the fact is that it falls significantly short of that claim. SIP is a phone protocol, and worse, one that has thus far been plagued with interoperability problems. Just as amazing, the mainstream press still refers to this 13-year-old protocol as an "emerging" technology and promotes it regularly. Should we not have made more progress by now?

We had hoped that the Internet would change the way we can communicate and we believe that it still can. What we do not believe is that there is any real hope that SIP will be at the center of that revolution. It might provide us with traditional voice or video communication capabilities, but that it about the extent of what we can expect. People will use it for now only because "it's what we have."

We can do better than that next time, and it is time we start looking forward.