Monday, April 25, 2011

Two technologies striking back in VoIP - part I

Living a Deja-Vu?!? Or maybe this time the dreams (of IETF/ITU/ETSI specs writers) come true ... who knows. This is a two parts story, about two technologies which many thought went dead long time ago but they are, again, actual in VoIP: IPv6 and IMS (Internet Multimedia Subsystem).

IPv6, the old story is on first pages again. Back in 2002 I worked at FhG Fokus Research Institute in Berlin, developing SIP Express Router (aka SER), within a majority of European Commission funded projects focusing on IPv6.

One may think this is not strictly related to VoIP, well, it is then at least very related to SIP, the leading routing protocol in VoIP these days. Why? Because SIP was actually designed for IPv6 networks. Why? Don't ask me.

Just slightly off-topic note, to refresh, for the past 10 years we had to invent, smash, re-invent solutions and do workarounds to get VoIP going through IPv4 networks -- or more specific saying, through NAT routers, since NAT was not considered in SIP specifications at all. Mixing STUN, TURN, COMEDIA and RTP relays, slipping on ICE, mangling headers and SDP, we got it working. Not few times, friends know that well, it had to be very specific hack per user agent or even firmware.

Back to IPv6, in Kamailio it has been IPv6 support to the core and main modules since May 2002, during the age of SIP Express Router v0.7.y, that's 9 years ago. Looking at the mailing list archive, it is now obvious the interest in the IPv6 is for the first time serious. Of course, the engine behind is the IPv4 address space shortage. Still there is no significant move towards IPv6 as reported recently. There are many (of us) not that happy to remember hex-based IP addresses, which may not be short at all, by the way.

If I hate (phone, digit-based) numbers, preferring instead SIP addresses (username@domain), think about the love I care for hex-numbers.

Moreover, fixing SIP for IPv4 somehow broke SIP for IPv6. Why? Because now we have to support all the nasty tricks specified to make SIP work for IPv4.

The IPv6 was started about a dozen years ago, initial battle was lost, now it is another one. Will it be the final battle? If IPv6 doesn't win this time, then the war is over for it.

To win, it requires fast and smart moves:
  • first and most important, is the adoption of IPv6 at the access level - core networks and home routes
  • then SIP devices (hard and softphones) with proper IPv6 support
If these two are not solved quickly, IPv4 shortage might be workarounded.

The last idea has a strong reason, non-technology related, but at the same time the biggest enemy for IPv6: the cost to switch to it. Internet is way bigger than 10 years ago, replacing/upgrading every piece running IPv4 now is going to be amazing expensive. Time was not the ally for IPv6 at all.

In the next post in this series it will be about IMS, yet another hot topic these days, might not be for VoIP, but surely is for VoLTE...

Maybe in the future we will get to the same situation for Enum? What about mobile IP?!?

No comments:

Post a Comment