[Iccrg] review of open-research draft

Eddy, Wesley M. (GRC-RCN0)[VZ] Wesley.M.Eddy at nasa.gov
Fri Sep 19 18:29:14 BST 2008


I've reviewed the open-research draft:
http://tools.ietf.org/html/draft-irtf-iccrg-welzl-congestion-control-ope
n-research-02

I found it to be much-improved over previous versions, and generally
in good shape.  I think the authors and editors have done a nice job
with it.  I know they still have a list of a few things they want to
do with it, but it seems near to publishable to me.  They've made some
very complex and thorny topics easy for a reader to quickly grasp and
understand the basics of.

Here are my comments:

- Section 1, 2nd paragraph: change "prima-dual" -> "primal-dual"
- Section 1, 3rd paragraph: at end of second sentence, a reference to
the
  ICCRG survey of congestion control RFCs may be sensible
  (draft-irtf-iccrg-cc-rfcs)
- Section 1, 4th paragraph: first sentence should have "in order to
avoid
  congestion collapse" appended to it.

- Section 2.1, first paragraph: change "interstellar" -> "space" ...
This
  one made me smile a bit because the Internet doesn't extend to other
  stars (yet) :). Getting it to other planets is proving to be
challenging
  enough that IP has been rejected by some parties (though not me :))!!!

- Section 2.1, 5th paragraph: This should also more clearly say that
multiple
  congestion control techniques interacting with one another is not yet
well-
  understood.

- Section 2.3, final paragraph: It seems to me that if an architecture
is
  built to support cost-fairness that equal-costs result in flow-rate
  fairness as a degenerate case; that is, flow-rate fairness is a
special
  case of cost-fairness.  One can be used to build the other, but not
  vice-versa.

- Section 3.7, it may be worth mentioning that its possible this issue
  would be different in a cost-based fairness regime, and it's worth
  thinking about how.  This section is interesting to me; I think the
  conclusion "By design, new congestion control protocols need to
  enable one end to check the other for protocol compliance." is pretty
  terrifying, and think that would be a cool topic to talk more about
  in the RG, personally.






More information about the Iccrg mailing list