2007 November

November 2007


In the stream of updated drafts for the IETF meeting in december, I found a draft namedReal-time text interworking between PSTN and IP networks.This is work-in-progress in Asterisk svn trunk. We already have support for T.140 realtime text in the SIP channel, and support for US TDD phones in the Zap/PSTN channel driver. During this summer, I’ve assisted a developer who implemented a RTT gateway in Asterisk, so you could call with an old PSTN TDD phone and have a conversation with a modern SIP client with support for realtime text.This code will soon be implemented in Asterisk svn trunk, so Asterisk will become the reference implementation for this draft. How many xOIP acronyms do we have now? FOIP, VOIP, TOIP? I surely missed a few there… :-)From the abstract:

IP networks can support real-time text communication. SIP-based real- time text is called Text-over-IP or ToIP. PSTN networks support real-time text using textphones (or TTYs). When real-timetext is supported by different networks, gateways are needed to provide interoperability. Real-time text capable gateways may also support real-time voice.This specification describes procedures for interworking between ToIP and PSTN textphones using a real-time text capable gateway (RTT gateway). It also describes ways to route calls to RTT gateways for several call scenarios. Procedures that support the phased introduction of RTT gateways and procedures that support the invocation of text channels at any time during the call are included. Interworking of PSTN text phones thatdo not support simultaneity of voice and text with IP User Agents that support simultaneous voice and text is also described.

The draft SIP Multiple REFER describes a way to send a call transfer (REFER) indicating setup of a new call to multiple destinations, i.e. initiating a conference. This could be handy between two Asterisk servers, a way to set up a conference on a different server, optimized for conferences.

RFC 3261 (SIP) [RFC3261] is extended by RFC 3515 [RFC3515] with aREFER method that allows a user agent to request a server to send a request to a third party. Still, a number of applications need torequest a server to initiate transactions towards a set of destinations. In one example, the moderator of a conference may wantthe conference server to send BYE requests to a group of participants. In another example, the same moderator may want the conference server to INVITE a set of new participants. We define an extension to the REFER method so that REFER request can be used to refer servers to multiple destinations. In addition, this mechanism uses the suppression of the REFER method implicit subscription specified in RFC 4488 to suppress REFER’s implicit subscription.

Josh Colp (file) has started a rewrite of the T.38 passthrough implementation in Asterisk. This is very much needed. The code that was contributed for 1.4 was a bad hack and has caused a large number of issues that Josh and I have been struggling with for a long time – and a large number of bug reports.

At the same time maybe we should sit down and consider the architecture of the SIP channel. After my initial planning on chan_sip3, I’ve worked a lot on video, text telephony (T.140) and other media in SIP. The SIP channel, and Asterisk, is built as a telephony audio-only PBX with some video additions. Adding T.38 (fax over IP), T.140 (text) and enhanced video/audio (like multichannel audio) turns into ugly hacks. And adding TCP has proven to be an ugly hack too. Yes, it’s easy to add a TCP socket, that’s no big issue coding-wise. But to handle the calls properly, the signalling and being able to switch from UDP to TCP midcall – that turns it ugly.

This also leads to a bigger question – what’s a PBX today? If we where about to start an Asterisk 2.0 project – would we handle all kinds of media properly? Presence? Messaging? File transfer? We already have an embedded web server… What’s the limit of a PBX? Or is it really a PBX we’re building? Asterisk 1.0 certainly has a PBX architecture, should that also be the base architecture for Asterisk 2.0?

Anyway, a big thank you to File for starting the re-write. May the code be with you!

We’ve all been there. You remembered you created an account on xxxyyyzzz.com, but
can’t remember the password or the account name. You search through your e-mail,
but can’t remember who sent you that confirmation mail. And you have enough of them.
That’s it. You’re note going to create yet another digital identity.

Identity on the net is an important issue. On one side, you got to be careful of the
electronic trail you leave. On the other hand, you need to make it easier on yourself
and limit the number of account names and passwords you use, so you end up
using the same password everywhere. I’ve reached the point that if I can’t
register the username ollej, my interest fades away… I don’t want help from my
web browser to remember, but I know a lot of people do. Changing to another
computer is a crisis situation for them. And letting the web browser handle
your accounts is not a very secure solution.

Identify is an important part of the security framework. It’s about claiming to
be someone (or something) and be able to prove it. This is very often called
AUTHENTICATION.

But that’s not all of it. There are different needs for how you prove your identity
in different situations. The requirement of a 100% correct identity is lower
on Facebook than some other sites, like your Internet bank. A few years ago
some people claimed that all secure transactions on the net required
200% security, being tied to your social security number. That was proven
itself to be incorrect – and you have to compare with daily life. How many
times have you asked your collegues for a passport and a DNA-test?

(more…)

BOB 2.0 - the conference.The world of realtime technologies on IP networks and the Internet is cooking with ideas, solutions and innovation. Combining it with social networks, and even more is happening. Protocols like SIP, XMPP/Jabber, SIMPLE, Jingle and Open Source solutions like Asterisk, OpenSER, Ekiga, Sofia-SIP, ReSIProcate and others are changing the landscape. It’s high time to get a helicopter view of what’s going on as well as detailed information about the building blocks for this new service delivery network.

Edvina.net is gathering leading experts in this area to the conference The IP Communications Technology Summit 2008 in Stockholm, April 2-4. Tutorials, hands-on labs, conference and user group meetings. Book this date in your almanac today! More information coming soon. If you have an idea or a talk proposal, send e-mail to me, Olle E. Johansson, at oej@edvina.net today!