Combined use of FNNs and AFNs

[Timothy I Mattox]

Hello,
We have been wanting to have the time to code up the combined use of a
high bandwidth network (such as a FNN) and our older, low latency AFN
technology, for quite some time.  However, the approach suggested for
using AFNs to enable advanced routing on FNNs is not appropriate.
As someone has pointed out, the time to load & unload modules, or to
dynamically change the channel bonding system would take too long to
be useful.  And, more importantly, the switches would not handle
the resulting changes in MAC addresses in their internal routing tables
quickly.
Advanced routing shouldn't need anything like that.  The difficulty
in doing FNN advanced routing is not technology, it is "simply a
matter of programming", which we have not yet had the time to do.
It is also a matter of determining at which layer to implement the
splitting of a data stream along multiple paths.  I personally see
several viable approaches, but none would work with the MAC slaving
system used in channel bonding, so it really has to be new code, not
just hacking on channel bonding in some way.  If anyone is interested
in discussing the technical details, we can discuss it off-list.

Some background for those who care:
Our research group has become rather small in our transistion from
Purdue University to the University of Kentucky, mainly to people
graduating, and getting "real" jobs. :)
At the moment, I'm the only Ph.D. student left working on developing
all theses things directly here at UofKY.  If you know of anyone who
would like to attend UofKY in the near future and work on this stuff with
us, please cosider it.  We have some really cool technologies waiting
for that eager student to flush out into reality.

P.S. - If you don't know what an AFN is, see: http://aggregate.org/AFN
They are Aggregate Function Networks, or "son-of-PAPERS" :)
--
Tim Mattox - tmattox at ieee.org - http://home.earthlink.net/~timattox
   http://aggregate.org/KAOS/ - http://advogato.org/person/tmattox/