[e2e] the evolution of deployability
ammar at cc.gatech.edu
Mon Jan 13 08:21:09 PST 2003
Jon Crowcroft wrote:
>and more heterogeneous - i wonder (and this was something that Mostafa
>Ammar, one of the panelists at ICNP talked about very nicely) if one
>could actually go so far as to form a proper (econometric, or systems
>science or perhaps even ecological) theory of network evolution? we
>must have other examples that someone could give us Bell, Hooke and
>Chandle on ?
For those who are interested, I have put my ICNP panel
presentation (from Nov. 2002) that Jon referred to
on line at
The slides are not quite the same without the words to go
with them so let me try to give some context.
The presentation was in response to the question of
what surprised us the most in the last ten
Among other things, I advocated the need to study
network evolution much in the same way that we study
the evolution of living beings. Along with this
comes three important points:
1- In order to study evolution we need to start
doing network "paleontology". This will help
us understand our history. Clearly this means that
our community needs to start putting a value
on looking back in time both in research and in
2- We need to develop evolution theories which
try to explain *why* the network evolved the way it did.
Note that much like the traditional science of evolution,
we need to keep the "what happened" and the "why it happened"
discussions separate. Hopefully, we can agree on the facts
of history but we can certainly debate the whys.
3- We need to foster healthier network evolution through
encouraging "bio-diversity" and "random mutation" in research
and through designing systems that make evolution relatively easy.
Regarding the design for evolution point, people may be interested
in looking at a contribution we made almost ten years
ago to the debate surrounding the design of IPng.
This is documented in RFC 1683 ("Multiprotocol Interoperability in IPng",
by Clark, Ammar and Calvert). Parts of the document
are somewhat dated now but the main argument of desiging
more flexible protocols and architectures seems
to be even more relevant today.
Mostafa Ammar Phone:(404)894-3292
Professor Fax: (404)385-0332
College of Computing E-mail: ammar at cc.gatech.edu
Georgia Institute of Technology
Atlanta, GA 30332
More information about the end2end-interest