Interoperability in SIP – state of the Art? – Part II

This is the second part of Interoperability in SIP and its relation to Art. You can read the first part here.

SIP Interoperability workshop

There were about 20 short but essential presentations at the 90 minutes workshop by the people representing different companies and organizations (yours truly represented IMTC and RADVISION). Overall message was rather unified. It is not all doom and gloom - there is basic interoperability in SIP. But all interoperability is extremely clustered - implementations usually work in the cluster of one vendor (sadly, if this is not always true). Henning Schulzrinne, for instance, was talking about the project done in the lab in Columbia University, where the SIP devices where acquired off the open market and then those devices run through the interoperability test in the lab - the success was, shall we say it, very limited. Similar experience was reported by lots of people in the audience.

What are the problem areas?

• Parsing is the first. There is a simple but often ignored rule - be strict with what you send, be lenient with what you receive. The sheer openness and ASCII encoding is literally encouraging people to add small pieces to the messages here and there - of course everything works in the cluster of devices from the single vendor - once deployed across interoperability becomes chancy.
• Another issue (rather global through) - is the time it takes to develop a standard. As SIP initially was focused on establishing simple connection over open internet with primary purpose of getting the audio through - this is what works best across domains and clusters. Simple audio call between various vendors has very good chances of success. As the standard was lagging behind actual multimedia communication market needs in development of advanced capabilities, the end result was that vendors decided not to wait for the standard and develop the functionality in the way necessary - and subsequently, the interoperability suffered greatly. Case in point - interoperability of real-time video implementations - extremely limited from vendor to vendor.

Well, let’s stop drilling on the issues - of course there is more. Now, as we are talking Art, and Art impact is usually positive, let’s look for the solution angle.

Improving SIP interoperability

What can be done to improve interoperability among rich multimedia communication SIP-based devices? There are a number of things which can be done.

1. For one, base interoperability profile definition would help. Interoperability profile reduces the number of things open for interpretation and enforces that compliant implementations will have higher rate of interoperability. Need an example? While gigantic overall, IMS is nothing but the specific (3GPP/TISPAN) profile developed on top of SIP. IMS-compliant systems would do much better in the real world (well, yes, come compliancy enforcement / certification program will be needed, but nevertheless).
2. Another helpful direction - implementation of the common reference architecture for deployment of any kind of rich media communications solutions in the network. ETSI is working in this direction, and IMTC is starting such work under the umbrella of Unified Communication and with help of Cisco leadership.
3. What else - further improvement of the test plans with the overall result/progress tracking during face-to-face and virtual IOT events really does help. This is how IMTC PSS and 3G-324M activity groups got so successful, and this is the direction of the IMS AG as well.

What else can be done? Well, I find it quite safe to predict (huh, predictions… in the tech world… yeah, right) that needs for and popularity of so called IP-to-IP gateway devices will be increasing in the near and mid-term future. Transgressing from one SIP to another SIP device, of from SIP to IMS or SIP to H.323 and so forth devices for that matter will be needed - and we will see more products of this kind on the market.

So, is it time for bottom line? Well, yes - I think we need to conclude this SIP and Art exploration. Is there a light on the horizon? Of course, by working together (SDOs, Consortia and Vendors) interoperability will improve. When? There is no hard timeline here - but it is important for the improvements to happen in the near term, otherwise the new forms of Art will emerge, and who knows, maybe they will be a bit easier to interpret, and therefore, embrace…