[conex] PhD dissertation relevant to ConEx audit

[Bob Briscoe <[e-mail]>]

Folks,

I (at last) have permission to publish my PhD dissertation.
Re-feedback: Freedom with Accountability for Causing Congestion in a 
Connectionless Internetwork
<http://www.bobbriscoe.net/projects/refb/#refb-dis>

It focused on what we now call the audit function in ConEx. Although 
it was based on the re-ECN protocol, I believe it is nearly all still 
relevant to the ConEx protocol.

It's 256 pages so not for the faint-hearted.
If you mainly interested in ConEx protocol desgn and audit, you only 
need to read ptIII (147pp) and it's double-spaced, so under 100 pages really!

It includes a wide range of potential attacks on the integrity of the 
protocol, which shaped the design of the protocol and of the audit 
function (at the time we called it the 'dropper'). It gives 
pseudocode for two alternative audit functions to maintain the 
protocol's integrity, including rationale and evaluation.

At least a couple of aspects have become dated:
* I have a much simpler but still as effective algo to replace the 
downstream congestion monitoring algo in S.8.2.7. I shall publish 
that as soon as I can - available on request.
* Mirja Kuehlewind has shown that the assumptions about TCP 
slow-start overshoot in S.7.7 were very over-optimistic


You will notice it acknowledges many who contributed. I would like to 
repeat my thanks to them all.




Cheers


Bob



________________________________________________________________
Bob Briscoe,                                BT Innovate & Design