camera_lumina at hotmail.com
Wed Jul 30 09:50:38 PDT 2003
This article appeared in the Lightreading telecom site. Although fairly
speculative, the mere implications of VoIP over P2P should be of much
interest to Cypherpunks, perhaps the least importance being encryption.
Ever since KaZaA, founder Niklas Zennstrom let it slip in a Boardwatch
interview that he was planning to launch a company offering voice services
using peer-to-peer (P2P) protocols, Light Reading and Boardwatch have been
puzzling over what he meant.
As Zennstrom himself has clammed up, speculation is in order. And our best
guess is that he might be onto something REALLY BIG. Think bigger than
Google. Think of something equivalent to the Internet's Domain Name System
(DNS) that would act as a distributed, dynamic, global telephone directory,
linking users with whatever IP address their appliance happened to be using
at the moment.
This addresses one of the big obstacles in the development of voice over IP
(VOIP), and it might also unlock other breakthroughs. For instance, it could
speed up the convergence of fixed and mobile telephony -- reckoned by Neil
Ransom, CTO of Alcatel SA (NYSE: ALA - message board; Paris: CGEP:PA), to be
the next big thing in telecom (look for the coming Light Reading interview).
It would be fairly straightforward to use P2P protocols to create such a
distributed, global telephone directory, according to Geoff Bennett,
director of Light Reading University, who happens to be moderating a Light
Reading Webinar titled "Controlling P2P: Who's Stealing Your Bandwidth?"
today, at 2:00 p.m. New York time. (Click here to register for the free,
With P2P, a user could search a directory using a special browser and then
click on the name he or she wanted to call. The client software would send
the request to the equivalent of a KaZaA supernode that would search a giant
routing table listing users and their current IP addresses, and send back
the result to the client so the VOIP call could be set up.
This is pretty much how file sharing works, and file sharing clearly scales
to the millions of users that might take advantage of such a system.
Zennstrom estimates that 100 million people already use P2P protocols, and
the KaZaA browser software has been downloaded a staggering 250 million
times -- an order of magnitude more than anything else on CNET's download
There are good reasons to believe that a P2P-based phone directory of this
sort might work better than the VOIP solutions being cooked up by the
Internet Engineering Task Force (IETF).
The IETF has various projects in this area, but they're all based on the
fundamental philosophy that existing technology -- namely DNS -- should be
adapted to deal with new requirements, rather than inventing something new.
But the new requirements in this case include more than just the IETF's
Telephone Number Mapping (enum) project.
The IETF has also got to come up with a way of making DNS more dynamic, so
that when a user moves from one IP address to another, he or she isn't cut
off from the world while the changes propagate through the DNS hierarchy.
When Boardwatch changed IP address earlier this year, a lot of readers were
cut off for two or three days. Imagine the same thing happening every time
someone shifted from a fixed to a mobile appliance!
Then there's the issue of reliability. Concerns are often expressed about
the security of DNS -- in particular, its vulnerability to denial-of-service
(DOS) attacks. A P2P approach promises to be much more reliable, because
it's so distributed and because the protocol has been designed to allow for
parts of the network going AWOL unexpectedly.
"Historically, the issue with proprietary solutions has been lack of
scaleability, and how robust they are in large networks," Bennett notes.
"But P2P apps like KaZaA have shown they can scale and have amazing
resilience because they're so distributed. How do you make a DOS attack on
something that big?
"The P2P solution will obviously be proprietary, and it's always quicker to
get a proprietary solution to market than to wait for consensus in a
standards group -- especially about something as institutionalized as DNS."
For more on IETF side of the story, check out the following columns by Geoff
Huston, a member of the Internet Architecture Board, the IETF steering body:
Lord of the Numbers
Who are You?
Light Reading is rather hoping that this speculation will encourage
Zennstrom to spill the beans on what he's up to. Right now, we think we've
only got half the story. The other half concerns SIP -- the Session
Initiation Protocol -- seen by some as another catalyst for a VOIP
explosion. In the Boardwatch interview Zennstrom slapped down the idea that
his project had anything to do with SIP. Which leads us to ask: Why was he
Peter Heywood, Founding Editor, Light Reading
The new MSN 8: smart spam protection and 2 months FREE*
More information about the cypherpunks-legacy