Disruptive Telephony

Just confirmed travel plans today - I will be heading out to the 70th meeting of the Internet Engineering Task Force (IETF) in Vancouver, British Columbia, Canada, from December 2-7. If any readers will be out there (either for the IETF or in Vancouver in general), please do drop a note and let me know. This will be my first meeting in my new role with Voxeo and I'm very much looking forward to renewing old acquaintances and also getting more directly involved with the work of the IETF.

Technorati Tags: ietf, sip, standards, voxeo

Did you know that RFC 4733 replaced/obsoleted RFC 2833? I just learned this myself through a SIP Forum mailing list exchange the other day. For those not aware, RFC 2833 and now 4733 define methods of carrying DTMF signals (and other similar signaling) in RTP streams separate from the main audio component of the RTP stream. A typical example of use might be where you were using a highly-compressed audio codec for audio between two SIP endpoints where the high degree of compression might make it challenging for the DTMF tones to be correctly interpreted on the receiving end. Using "RFC 2833 compliant" signaling, the sending SIP endpoint would send those DTMF tones as separate packets within the RTP stream.

My key takeaway from learning about RFC 4733 is that we should really be talking about "RFC 4733 compliant" signaling... but given that the industry is really only now starting to really talk about "RFC 2822 compliant" signaling, I'm not sure I expect to see that happening anytime soon.

Anyway, here's the abstract from RFC 4733 - you can naturally read the rest of the document to understand more:

This memo describes how to carry dual-tone multifrequency (DTMF) signalling, other tone signals, and telephony events in RTP packets. It obsoletes RFC 2833.

This memo captures and expands upon the basic framework defined in RFC 2833, but retains only the most basic event codes. It sets up an IANA registry to which other event code assignments may be added. Companion documents add event codes to this registry relating to modem, fax, text telephony, and channel-associated signalling events. The remainder of the event codes defined in RFC 2833 are conditionally reserved in case other documents revive their use.

This document provides a number of clarifications to the original document. However, it specifically differs from RFC 2833 by removing the requirement that all compliant implementations support the DTMF events. Instead, compliant implementations taking part in out-of-band negotiations of media stream content indicate what events they support. This memo adds three new procedures to the RFC 2833 framework: subdivision of long events into segments, reporting of multiple events in a single packet, and the concept and reporting of state events.

Technorati Tags: DTMF, ietf, sip, standards

In the usual (and ongoing) flurry of IETF announcements, there was one notice that caught my attention.  It announces that an Internet Draft document about "dialstrings" has been approved to become a standards-track RFC.  So what, you say?  Well here's a bit more info:

This document provides a way of incorporating a dial string into the SIP or SIPS URI scheme. A dial string is a cousin of a telephone number, but rather than taking the form of a fully-qualified E.164 or national-specific telephone number, it is a description of a literal set of dialed digits that would be delivered over a POTS line. As such, it may include pauses, omit prefixes like area codes, and its applicability is necessarily restricted to a particular context (an enterprise, a LATA, etc). Support for dialstrings was formerly a feature of the tel: URI scheme specification (back in RFC2806); since that functionality did not make it into the revision (RFC3966), it is provided here specifically for the SIP and SIPS case.

Think of it as extra digits you have to type when making a call... or extra keys you have to press to start a service.  The challenge is that SIP proxies and other services need to know that it is a string of numbers that should be handled in a special manner, rather than just thought of as part of a SIP address or something like that.  I mention it here only because it's one of those really low-level things that you can do on the PSTN but until now haven't had a (standard) way to do in SIP. It's also one that ultimately anyone implementing SIP will need to implement.  No RFC number yet, but that will come soon.  Note the nice security warning:

Dialstrings exposed to the Internet may reveal information about internal network details or service invocations that could allow attackers to use the PSTN or the Internet to attack such internal systems. Dialstrings normally SHOULD NOT be sent beyond the domain of the UAC. If they are sent across the Internet, they SHOULD be protected against eavesdropping with TLS per the procedures in [RFC3261].

Yep... as we've been saying over at VOIPSA and Blue Box, you definitely need to think about encrypting SIP if you are sending it across the Internet.  If not, bad things will happen eventually.

 So you'd like your SIP phones to all work together, eh?  And you'd like your SIP phone from Vendor A to work with the SIP phone of Vendor B and yet give you the business functionality that you used to have in the PBX from Vendor C?

Good luck.

Yes, they will (or should!) all work together for basic call functions, but if you want to do more than just the very basics, you rapidly wind up in the realm of incompatible SIP implementations.  Different vendors support different RFCs... or interpret RFCs differently.  It's a challenge to go beyond basic functionality.

Enter "BLISS", one of the latest working groups coming out of  the IETF. It stands for "Basic Level of Interoperability for SIP Services" and, as noted in its charter, the intent is to define a basic set of functionality ("minimum interoperability requirements") to allow SIP endpoints to interoperate on 4 specific telephony services:

1. Bridged/Shared Line Appearance (BLA/SLA)
2. Call Park/Pickup
3. Do Not Disturb (DND)
4. Call Completion to Busy Signal/Call Completion on No Reply

More details are on the charter page.  These are just the initial four issues chosen to be addressed and Internet-Draft documents are already circulating on some of the items.  I see it as a necessary group if we are to actually have real interoperability between SIP endpoints, and I commend the group organizers on scoping it to an initial 4 issues (rather than making it wide open).

If you work with SIP, I'd encourage you to check out the BLISS web site, read the "problem statement" and charter... and then join the mailing list (and please read the archives to see what's happened so far).  I've joined the list... and would encourage others to consider doing so.  If we do want real SIP interoperability, we need to hammer some of these things out.

For more information, you can download the set of slides presented by Jonathan Rosenberg of Cisco at the IETF 68 meeting stating the problem and why the BLISS working group is needed.  To make it easy to see them, I'm going to embed a SlideShare version here:

See you on the list...

So with the release yesterday of RFC4824, The Transmission of IP Datagrams over the Semaphore Flag Signaling System (SFSS), one has to wonder... which of the vendors will be the first to attempt to implement VoIP transmission in this medium?     I think it would make for a rather slower conversation, but it would certainly be intriguing.  Hmmm... I wonder which would be faster - this method?  Or the avian method defined in RFC2549.  Probably this one, methinks. 

Oh, you have to love a standards body with a sense of humor...