General Information:
Instructors: Katia Obraczka and Abdul Kabbani
Classroom: BE 169
Class time: Tue 3:20-4:55
Class Description:
This class will cover topics on cloud computing and datacenter networking based on recent research papers. We will read and discuss one paper per week. One student will be responsible for presenting the paper to the class and leading discussion. Each student should submit a 1-page critique of the paper including answers to the following questions: what is the problem the authors are trying to solve? why is the problem interesting, relevant, and/or important? what other approaches or solutions existed at the time that this work was done? how the proposed approach contributes to the state-of-the-art? what is the proposed approach and how does it compare to earlier approaches? what are the main strengths and weaknesses of the paper/proposed approach? The presenter does not need to submit the critique; instead they should submit their slides beforehand which will be posted on the clas Web page. Presentations should be designed to provide a brief overview of the paper and foster discussion.
Tentative Schedule
| Date | Topic | Presenter |
| April 2 | Course overview and logistics | |
| April 9 | Congestion control | Chris |
| April 16 | Minmei's dry-run (to be confirmed) | Minmei |
| April 23 |
Datacenter network architecture and topology design |
Sam |
| April 30 | Network telemetry | Daniel |
| May 7 | Load balancing and traffic engineering | Li |
| May 14 | Space and interplanetary internetworking | Kerry |
| May 21 | Machine learning and networking | Andrea |
| May 28 | Yalda's dry-run | Yalda |
| June 4 | SDN and datacenter networking | Dylan |
Reading List:
Congestion Control
1- Re-architecting datacenter networks and stacks for low latency and high performance Sigcomm 2017
2- Copa: Practical Delay-Based Congestion Control for the Internet NSDI 2018
3- Revisiting Network Support for RDMA Sigcomm 2018
Public-key certificates for IoT (Minmei's practice talk)
Collaborative Validation of Public-Key Certificatesfor IoT by Distributed Caching, Infocom 2019
Network Architecture and Topology Design
1- Understanding Lifecycle Management Complexity of Datacenter Topologies NSDI 2019
2- B4 and After: Managing Hierarchy, Partitioning, and Asymmetry for Availability and Scale in Google’s Software-Defined WAN Sigcomm 2018
3- RotorNet: A Scalable, Low-complexity, Optical Datacenter Network Sigcomm 2017
Telemetry
1- Inferring Persistent Interdomain Congestion Sigcomm 2018
2- NetBouncer: Active Device and Link Failure Localization in Data Center Networks NSDI 2019
3- Is advance knowledge of flow sizes a plausible assumption? NSDI 2019
Load Balancing
1- Resilient Datacenter Load Balancing in the Wild Sigcomm 2017
2- DRILL: Micro Load Balancing forLow-latency Data Center Networks Sigcomm 2017
3- Semi-Oblivious Traffic Engineering: The Road Not Taken NSDI 2018
Space Networking
1- Delay is Not an Option: Low Latency Routing in Space
2- Gearing up for the 21st century space race Hotnets 2018
Machine Learning and Networking
1- Resource Management with Deep Reinforcement Learning
2 - An Experimental Study of the Learnability of Congestion Control
3 - TCP ex Machina: Computer-Generated Congestion Control
Yalda's dissertation defense dry-run
- Yalda Edalat, Katia Obraczka, and Bahadur Amiri, “A Machine Learning Approach for Dynamic Control of RTS/CTS in WLANs”, in IEEE Mobiquitous 2018. pdf
- Y. Edalat, J. Ahn and K. Obraczka, “Network State Estimation Using Smart Experts”, 11th International Conference in Mobile and Ubiquitous Systems (MOBIQUITOUS), London 2014.
- Edalat, Yalda, Jong-Suk Ahn, and Katia Obraczka. “Estimating Network Performance Using SmartExperts” DOI 10.1109/TNSM.2016.2586506, IEEE Transactions on Network and Service Management.
SDN and datacenter networking
- B4 and After: Managing Hierarchy, Partitioning, and Asymmetry for Availability and Scale in Google’s Software-Defined WAN Sigcomm 2018