Reading Review for Class 7
This class’s readings are:
Optional:
Your review should include both a summary of these papers, which
discusses
- the problem the paper addresses
- the paper’s solution
- the evidence the paper provides that the solution was
correct. Remember to keep in mind the historical context the
papers discuss.
Your review should also discuss, after the summary:
- whether the problem the paper solved is an important
one
- whether the solution was a good one
- whether the paper was fun or enjoyable to read.
- what questions you have about the paper, or which parts of the paper
(if any) you found hard to understand.
You should answer the above questions in about one sentence each. The
remainder of your review (about the below questions) should be about a
page (~500 words) long.
Finally, for this week’s papers, please reflect on the following
questions:
- Here
is an implementation of deficit round robin, a fair queueing algorithm
described in the DRR optional reading.
- This implementation won’t provide perfect fairness between flows.
What factors might cause it to provide less-than-perfect fairness, and
how unfair might the allocations be?
- How might you modify this implementation to prioritize different
flows or groups of flows (describe in text, code optional)? How would
the resulting algorithm compare in complexity and performance to the
HDWRR algorithm in Appendix A (also here)?
- Would Briscoe ’07 likely agree with this proposal or not, and why?
Would Briscoe ’07 likely prefer RCS, or the “Network Utility
Maximization” approach described briefly in Section 8 (or neither, or
both)?