Topicality perceptions

Alexander Harrowell a.harrowell at gmail.com
Mon Sep 25 10:37:27 UTC 2006


Concur. Nanog has been an on-going education in essentially all
aspects of internetworking, routing, data centres, security,
spam/malware/abuse. Long may it stay that way. I'd argue that the
fuzziness is probably a reflection of the ever-broadening role of
IT/telco/netops people and ideas in current organisations.

Now, someone mentioned issues with SIP. I'd like to flag that this is
going to become a top line operational issue in the next few years,
due to the deployment of following technologies:

1) Carrier/Enterprise VoIP
2) Peer-to-Peer VoIP using SIP (see - Gizmo)
3) Concurrent applications using SIP
4) IP Multimedia Subsystem (IMS) in mobile networks (and possibly
fixed networks) interworking with each other, PSTN and the public
Internet
5) ETSI TISPAN activity (probably the least important of the five)

Note that 1 through 3 use SIP as defined by IETF whereas 4 and 5 use
the 3GPP/3GPP2/ETSI "extensions" to it, which may mean they cannot
interwork. Further, IMS and various associated technologies employ DNS
ENUM to map e164 numbers to SIP URIs, not to speak of ordinary DNS to
map URIs to IP addresses.

Some DNS security measures previously discussed on NANOG have the
effect of filtering ENUM replies. There is also the problem that IMS
carriers, as far as anyone knows, are going to operate as private
internetworks and do some form of NAT at the Session Border Controller
(ie - gateway to the public Internet). How they will handle this at
private interconnections with each other is unclear. It is also
unclear how connections between a "Carrier SIP" client with a
privately assigned or RFC1918 address and a carrier-land URI, and an
open-Internet "IETF SIP" client with a globally routable address and
its own URI, will work.

It also seems clear that IMS-adopting carriers will continue to
declare themselves as carrier grade, which suggests that the
criticality of their private DNS will be very high.



More information about the NANOG mailing list