R02 ACS MPhil - Network Architectures
Second Essay (Set Nov 13, due Dec 6)

This essay is about ML and Networks. There are two fairly recent papers which show how AI techniques can be applied to
adapting congestion control algorithms; and can be used to provide robust traffic engineering.

What is interesting about these papers is that the authors do not just apply a hammer to a nail, but rather, subtly unpick
the problem both from the networking perspective, and from the appropriate and proportionate use of AI to the particular
problem under study.

This essay should be about that tradeoff in design in each case, and you should try to see if we can draw more general
lessons about the use of AI and Machine Learning ideas in particular application problem domains.

Background to ML&CC
PCC Vivace: Online-Learning Congestion Control
https://www.cl.cam.ac.uk/teaching/2425/R02/papers/nsdi18-dong.pdf 

Background to ML&TE
DOTE: Rethinking (Predictive) WAN Traffic Engineering
https://www.cl.cam.ac.uk/teaching/2425/R02/papers/DOTE.pdf 

In the case of the DOTE paper, it isn't necessary to discuss the material in the appendices
(unless you want to).


As per the general rules, essays are 2-4 pages, should cite all sources,
and can be in note form - the use of tables (for comparisons) and
diagrams (for clarity) is encouraged.