We welcome and encourage reader response to stories and commentary in our magazines and on our sites. Here is a letter received from Paul E. Jones Chair, H.323 and H.325 Experts' Groups, regarding a recent article on SIP

Editor, Customer Interaction Solutions

I felt compelled to comment on the second page of your most recent article: S(L)Ipping IP Into The Contact Center (May 2009) http://flq.us/uZ
 
"[SIP] is gradually displacing H.323, which is an older protocol used to deliver voice over an IP network.

* They are the same age
* H.323 was designed for videoconferencing and multimedia, not "voice"
 
"SIP has several advantages over H.323, including scalability, ease of integration, and ease of management."

* Scalability is debatable, as we can build a massively scalable H.323 network, and we have

* "Ease of integration"...with what?  Perhaps it might be easier to integrate a half-baked SIP solution with another half-baked solution?  A fully-robust solution is a lot of work, regardless of protocol

* Ease of management?  Apparently the person who made this claim is entirely ignorant of the capability negotiation issues with SIP, the fact there is no means of accurately directing calls toward gateway devices to ensure calls do not fail on overloads, etc.  Poor management is actually something H.323 and SIP have in common, but there is absolutely nothing better in SIP.
 
"SIP has also helped lower product costs because it has made the components more interoperable through the common standard. Tom Fisher, Director of Systems Engineering for Interactive Intelligence points to an IP endpoint written to H.323 in 2001 that sold for $900, which has been superseded by one done in SIP in 2008 for $120."

* H.323 and SIP phones made in China run on exactly the same cheap hardware

* There were H.323 endpoints in 2001 that retailed for less than $150, though videoconferencing equipment was generally more expensive; in general, videoconferencing equipment is more expensive due to hardware costs that have nothing to do with the protocol employed
 
What bothers me is that SIP and H.323 are both now 13 years old (counting from the initial date of draft publication).  Even if they were a couple of years apart, the fact remains that SIP is "ancient" by all accounts.  Just how old was the "old" SS7 network when SIP was introduced in draft form in February 1996?  I believe it was 16 years old, though I might be mistaken.
 
The point is that, while SIP is a workable protocol for VoIP applications, it has a lot of problems.  It is better in some respects than H.323, but is worse in some respects than H.323.  To this point, SIP has done very little to propel the market forward: what propelled the market was the notion of VoIP, accelerated by H.323, and now slowly progressing forward with a mix of SIP and H.323 being deployed.  While it is true that there is a slow migration to SIP-based networks, most deployments have been little islands, not unlike the little H.323 islands.  Carriers around the world are trying to pull together some sort of interoperable SIP-based network (which they have failed to do thus far due to equipment interoperability issues and missing "stuff" to make carrier networks robust).  But the sad thing is that what they will deliver when they're done is nothing more than the PSTN over IP.  Have you looked at IMS?  It's a complex mess.
 
The big question is: do we need that?  Look at Skype.  It works for me, and they are basically a global "carrier".  One can put voice on just about any device and various companies can provide interworking through SBC devices in the back-end.  Perhaps it's SIP, or perhaps it's proprietary glue.  It matters very little to the end user, since a "call is a call".
 
What is worth looking out for is a new way of communicating, like being able to use a plurality of devices in parallel.  For example, wouldn't it be nice to use a headset to place a call (perhaps connected to a phone), the phone is talking to a mobile device in your hand.  The mobile device is also talking to your PC, so when you join a conference bridge, a notification is transmitted to your PC to let you join a web collaboration session: all without having to do a thing, really.  Oh, and during the call, perhaps you can transmit a file from your PC to the person(s) you're talking to, because there is an app on your PC that knows about your active call and transmits the file as a part of the session.  Now, that's cool.  But, it's not SIP.  It's a new kind of XML-based protocol that we're working on at the ITU right now called H.325 (or, "AMS", not to be confused with IMS).
 
I have no idea whether H.325 will ever prove to be a market success, but I would really like to see articles that do not continue to convey the same misleading (and I doubt it's deliberately misleading) statements.  The world really is not better with SIP.  In the words of an associate from in the VoIP industry, "SIP sucks.  It really sucks.  But, it's what we have."  And, there's a lot of truth in those words.

Don't let my comments suggest that SIP will fail. I suspect it will succeed. After all, what other options are there? Enterprises need something and carriers need something. But, SIP has long been over-sold as something that will enable all kinds of new applications. I strongly believe in the power of IP-based multimedia communications and SIP will play an important role; it's just over-sold and has been for so many years.

Yours truly,

Paul E. Jones
Chair, H.323 and H.325 Experts' Groups