Where in the World is H.323?

By Tsahi Levent-Levi

Categories: Standardization

February 18th, 2010

Earlier this week I’ve outlined what the SIP parity group at the IMTC is all about, but then again – a question does arise – where in the world is H.323 these days?

Does Carmen Sandiego know where H.323 is? (image by ~5ir3ntropy)

I’ve heard people announce the death of H.323 a lot of times. I guess the first time was about 8 years ago. Funny thing is, it’s still going strong – we actually have new customers who license our H.323 stack.

So the question is why? Why go to all the trouble of licensing and then adding H.323 to a product at this point in the game, when surely H.323 has lost. If I said a year ago, that H.323 is the only thing that is used for corporate video communications, now – with the work done by the IMTC – that too is lost.

I’d say that H.323 is still here because technology isn’t adopted as fast as we think, and when it does it is either because it makes too much sense for the vendors who are developing the relevant products (mainly because of cost reduction issues or increased flexibility) or because it makes too much sense for the end users.

Need a few examples?

• Vendors: A lot of the calls we make today internationally (or even locally through the cable operators) are done using VoIP. We, the customers, don’t care about it, but it allows operators to cut their costs – for us it’s the “same” service, for them it’s a whole different technology.
• Users: Mobile data plans were never the rage. That is until Apple came out with the iPhone and a new way to surf the mobile web. So only now operators are really trying to figure out how to deal with the traffic load.

And where does that leave H.323 and SIP? The IPv4 vs. IPv6 bickering comes to mind. From the day I started working in RADVISION (more than a decade ago) I’ve been hearing that it’s mandatory for our customers to use IPv6, because this is what their customers ask for. Guess what? Most of our customers just have it on the feature list but don’t really need it. IPv6 is like a solution looking for a problem – and don’t talk to me about the depleting availability of public addresses or the multiple NAT traversal issues – we’ve been living with those for a whole decade being feared and terrorized by technologists.

IPv6 will end up happening, but a lot slower than technologists want.

So H.323 is just like IPv4 in that sense. It will fade away, but a lot slower than SIP enthusiasts want. The last stronghold of H.323 is definitely the enterprise video conferencing market, as most deployments today are H.323 deployments, and SIP is only used on the fringes of this market. In this context, SIP is a solution looking for a problem, as nothing that SIP can do – H.323 can’t.

But there is a change – SIP is going to win eventually, and the reason is that as video use grows and more people adopt it as one of their communication means, it has to find its way into our mainstream communication solutions – it has to connect directly and more intimately with the PBX. And PBXs today are ruled by SIP.

The move to SIP won’t come from the users – they couldn’t care less. It will come from the vendors who need to connect it to PBXs.

But it won’t happen today. The SIP parity group is a piece in the puzzle and a move in the right direction. But the way is still long and arduous.