TcpEvalProject

ICCRG TCP Evaluation Suite port - Summer Project 2016

• Student: Dharmendra Kumar Mishra
• Mentors: Mohit Tahiliani , Tom Henderson
• Abstract: A lot of new TCP extensions have been proposed over past few decades. Comparing the performance of these TCP variants is a tough job due to a lack of agreed set of performance metrics. To address this problem, the Internet Congestion Control Research Group (ICCRG) has provided a common test suite for evaluating new TCP extensions. This suite is already available for ns-2. The goal of this project is to add Delay/Throughput tradeoff scenario [section 5.1 of the RFC] to the existing code of TCP Evaluation Suite for ns-3. This project, which is a port of the ns-2 version suite, also aims to modify the existing ns-3 TcpEval code to make it compliant with the RFC.

• Referred RFC: Common TCP Evaluation Suite [ https://tools.ietf.org/html/draft-irtf-iccrg-tcpeval-01 ]
• Referred ns-2 code: [ https://bitbucket.org/hayesd/tcp-evaluation-suite-public ]
• ns-3 TcpEval code: [ https://github.com/dharmendra-mishra/TcpEvalSummerProject ]
• About me: M.Tech in Computer Science from National Institute of Technology Karnataka, Surathkal, India. Currently, I am hired as Network Consultant by CISCO Systems, Bangalore, India.

Weekly Plan

Week 1

* Identify the changes (topology setup, parameter configurations) required to simulate Delay_Throughput_Tradeoff 
  scenario in ns-3 (section 5.1 in RFC)  
* Implement the changes.
* Document them in a .rst with reference to the parts completed in the RFC

Week 2

* Test/review Tmix integration with ns-3.
* Generate traffic in ns-3 TcpEval using Tmix.
* Update the documentation.

Week 3

* Write code to shuffle the Tmix traces (section 2.3.2 in RFC)
* Test the working of Tmix module against the traces used by ns-2 TcpEval

Week 4

* Implement the statistics generation in ns-3 TcpEval by using ns-3 callback based tracing.
* Following metrics (section 4.5 in RFC) for the central link will be calculated:
  1. the aggregate link utilization,
  2. the average packet drop rate, and
  3. the average queueing delay.
* Document the working.

Week 5

* Continue week 4 work: per flow statistics generation

Week 6

* Compare the results obtained from ns-2 TcpEval and ns-3 TcpEval for Delay_Throughput_Tradeoff scenario 
  (section 5.1 in RFC).
* Document the observations (similarities and differences)
* Mid-term review can be scheduled around this week.

  In the project, following issues may consume more time:-
    1. Integration of Tmix with TcpEval for ns-3.
    2. Shuffling of Tmix traces.
    3. Per flow statistics generation.

Week 7

* Continuation of Week 6: conduct detailed comparison in different configurations of Delay_Throughput_Tradeoff 
  scenario (section 5.1 in RFC).
* Document the observations
* Write a new test suite for Delay_Throughput_Tradeoff scenario (section 5.1 in RFC).

Week 8

* Respond to comments received on the code review.
* Finalize any remaining issues before declaring the code to be 'ready to merge' state.
* Finalize the documentation.

Week 9 to 12

* Simulate RED or other qdiscs using the ns-3 TcpEval implemented in previous 8 weeks.
* Compare the results with ns-2 TcpEval.
* Test and document the observations.