CS 268 (Adv Network)


How to read a paper

Reading paper

  • Sylvia's advice: do not overthink it, just read it from start to end in depth

  • Phase 1: Understanding --> Questions to answer

    • What problems are they solving?

      • In 1-2 sentences

    • Replay their motivating arguments

      • Do they make sense

    • Replay how their solution works for a simple example

      • Can you? If not, go back to step 1

  • Phase 2: Critiquing

    • Is the problem important? ambitious? hard? having a long-shelf-life?

    • Is the solution effective? Under what conditions does it break?

    • What other approaches are possible?

Reviewing paper

  • Review is not a summary

  • A typical format of the review

    • What is the problem being addressed?

    • Do you believe that the problem is / was important?

      • Explain your thinking

      • Consider context

    • What is the solution's main insight (nugget)?

    • Do you think the solution is a good one? (strengths and weakness) --> explain

    • Did you enjoy the paper?

Reading list

Intro and course overview

Robert E. Kahn and Vinton Cerf 2004 ACM A.M. Turing Award Lecture

Internet Architecture

Beyond best-effort / Unicast

Congestion Control

Datacenter Networking

SDN: context

SDN: practice


BGP Security

Programmable Networks

Datacenter Congestion Control

WAN Congestion Control

P2P Networking

Net sw



Low latency


Last updated