Preparation for Bufferbloat recitation
Read both the paper by Jim Gettys and the paper by Mark Allman (it will be helpful to read Gettys' paper before Allman's, as it contains some background information about TCP). Both of these papers study the same topic---bufferbloat---but draw different conclusions.
As you read, you should come to understand some basic facts about bufferbloat:
- Why do queues exist in the Internet?
- What is bufferbloat?
- How does drop-tail queueing affect bufferbloat?
- How does bufferbloat affect TCP specifically?
- How does bufferbloat affect small flows (compared to large flows)?
You can also check out Question 8 from the 2014 quiz (solutions).
As you read, think about the following:
- Both papers reference a third paper, about a system called Netalyzr (Kreibich, et al. "Netalyzr: Illuminating the Edge Network". IMC 2010). What conclusion(s) does Gettys draw from the Netalyzr paper? What conclusion(s) does Allman draw?
- What types of experiments does Gettys do? What types of experiments does Allman do?
- Do you believe that bufferbloat occurs frequently on the Internet?
Question for Recitation
Before you come to this recitation, write up (on paper) a brief answer to the following (really—we don't need more than a couple sentences for each question). If your TA has requested that you email your answer to them, you may do that instead, but it should still be handed in before your recitation begins.
Your answers to these questions should be in your own words, not direct quotations from the paper.
- What conclusions does each author draw about bufferbloat?
- How do they support those conclusions?
- Why do the two authors draw different conclusions in some cases?