ACM SIGCOMM 2024 Non-Paper Sessions
5 August
Can ML models replace network algorithms? Feasibility, challenges, and how to get there
Organizers: Mohammad Alizadeh (MIT), Behnaz Arzani (Microsoft), Bruno Ribeiro (Purdue), Keith Winstein (Stanford), Sanjay Rao (Purdue)
In recent years, we have seen rapid growth and impressive initial results in the use of ML models in lieu of classical network algorithms in domains such as video streaming, Internet routing, traffic engineering, video coding and data-center networking. While this offers exciting opportunities, networking environments pose many unique challenges for ML approaches. In networking, we are rarely content with just observing a system evolve: it is common to intervene (e.g., adapt bit rates, reroute traffic, switch CDNs) in order to improve it. Unfortunately, online experimentation in live systems is not always viable, and can be expensive. Unobserved confounders (e.g., intrinsic network conditions, queuing policy) may significantly influence the effect of an intervention. Moreover, the predicted output from an ML model must extrapolate to settings and actions not present in the training data (e.g., unseen failures, traffic surges). Many directions at the cutting-edge of ML research can help address these questions, yet domain-inspired insights are crucial, and adapting them to the networking context poses unique challenges.
The role of industry and academia in network research
Organizers: Zili Meng, Hong Kong University of Science and Technology
Industry and academia are two extremely important participants in the network research. For decades, both industry practitioners and academic investigators have contributed numerous intellectual results to the community. Network research has always been a good example on how the collaboration between industry and academia can be beneficial to both. There has been a recent trend of more industry-oriented papers in top conferences such as SIGCOMM/NSDI. For example, the ratio of papers with at least one industry affiliation in SIGCOMM/NSDI is around 50% now. SIGCOMM has respectively set up a separate experience track since 2020 as well. On one side, some academic researchers will worry that it'll be difficult to make research contributions without the evaluation over large-scale production-level deployments, given the scale of a typical research group. On the other side, seeing more experience papers from the industry is also a valuable opportunity for researchers to understand how the real network systems work in production.
SIGCOMM reviewing Q&A
Organizers: Balakrishnan Chandrasekaran, Vrije Universiteit Amsterdam, Yiting Xia, Max Planck Institute for Informatics
Effective paper reviewing is paramount to fostering the organic growth of the SIGCOMM community. As the landscape of networking research continues to evolve, esteemed members of the SIGCOMM community have underscored the importance of enhancing the inclusion and transparency within the reviewing process. However, newcomers to our community, including students and researchers with backgrounds in other domains, still likely perceive the reviewing process as opaque and elusive. In response to this need for greater clarity and inclusivity, we propose the organization of a dedicated workshop session. This session aims to achieve two primary objectives: 1) Facilitating Dialogue: We intend to provide a platform for Technical Program Committee (TPC) chairs and committee members to engage with a diverse cross-section of our community. Through open dialogue and feedback collection, we seek to gain insights into the specific needs and expectations of authors, thereby fostering a deeper understanding of the evolving landscape of research contributions. 2) Knowledge Dissemination: Furthermore, the workshop session will offer an opportunity for paper authors and students to gain invaluable insights into the criteria and processes involved in the evaluation of research papers. By elucidating the evaluation criteria and demystifying the reviewing process, we aim to empower participants to navigate the submission and review process with greater confidence and clarity. Through these concerted efforts, we aspire to catalyze a culture of transparency, collaboration, and inclusivity within the SIGCOMM community, thereby ensuring the continued vitality and relevance of our research endeavors.
6 August
SIGCOMM Debates
Organizers: Akshay Narayan, Brown University
SIGCOMM should include a non-paper session centered on one or more moderated technical debates on specific networking-related topics. The SIGCOMM debate session will present two moderated technical debates on the following topics: 1. On community deployability standards for congestion control algorithms. Ayush Mishra (NUS) and Aditya Philip (CMU). 2. On the role of LLMs in networking research. Behnaz Arzani (MSR) and Keith Winstein (Stanford).
Computer Networks Education in the New Era: Challenges, Opportunities, and What Should SIGCOMM Do?
Organizers: Qiao Xiang, Xiamen University
Every year, we SIGCOMMers organize many research conferences, workshops, and seminars to share new research ideas and results. However, there are few such venues to talk about education issues in computer networks. This side event proposes to discuss (1) the challenges and opportunities of computer network education, at both undergraduate and graduate levels, in the new era of emerging technologies (e.g., AI, quantum, and new computing architectures), and (2) what should SIGCOMM do in this new era to better prepare students with sufficient competencies in computer network for their future career.
7 August
Reflections on the research/product design process
Organizers: Vyas Sekar, CMU
In this "meta" talk, we will cover the often intangible part of the research and design process in how we can be more effective in problem formulation, choosing problems to work on, devising strategies to tackle problems the right way, and how we can be more effective in communicating our results. This is the sort of talk I wish I had heard 20+ years ago when I was starting out as a junior researcher.
SIGCOMM from the Past: Stories & Experiences
Organizers: Behnaz Arzani, Microsoft Research
Personal storytelling has recently become a popular hobby for many (the Moth and Fresh Ground stories are two popular story telling venues). The topic of the story-telling session at SIGCOMM is to hear stories about what people experienced in their first SIGCOMM, how they perceive SIGCOMMs evolution through the time they have been part of it, and what are interesting stories that have happened to them at the conference.