CS 244 ’19: Reproducing Fidelity and Scalability of Congestion Control Plane Algorithms

Download report

Robert Neff, Kevin Rosendahl

Original paper: Akshay Narayan, Frank Cangialosi, Deepti Raghavan, Prateesh Goyal, Srinivas Narayana, Radhika Mittal, Mohammad Alizadeh, and Hari Balakrishnan. 2018. Restructuring Endpoint Congestion Control. In Proceedings of the Applied Networking Research Workshop (ANRW ’18). ACM, New York, NY, USA, 25-25. DOI: https://doi.org/10.1145/3232755.3232770

Congestion control algorithms implementation and research continues to be important area of study as new ideas require extensive testing on increasing numbers of application- network stacks or “datapaths” and seek to provide different functionality. This report presents our reproduction of Restructuring Endpoint Congestion Control by Narayan et al. which explores placing algorithms in a user-space agent outside the datapath. In recreating Figure 10 we find that algorithms run in this mode match congestion window, throughput and delay fidelity of their kernel counterparts fairly accurately under a variety of simulated network conditions. We further replicate Figure 13, noting that multi-flow throughput and CPU utilization scale according to the paper and find that the additional time overhead costs are dominated by context switches. Finally, we briefly attempt to implement TCP Reno using their code for datapath setup [8].

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s