ACM SIGCOMM 2025 Non-Paper Session
Research Infrastructures and Tools for Reproducible Data Communication Research
Date, Time and Room
- Topic Research Infrastructures and Tools for Reproducible Data Communication Research
- Where/When Tuesday, September 9th, 17:30 — 18:15, Room Mondego.
- SIGCOMM website information
Organizers
- Georg Carle (Technical University of Munich)
- Serge Fdida (Sorbonne University)
- Henning Schulzrinne (Columbia University)
Motivation and Goals
Experimental research on networked systems requires suitable research infrastructures (RIs). For many years, the dominating approach of research groups was to create a specific experimental setup tailored to the needs of a specific set of experiments, e.g., in the context of a PhD thesis. The community is well aware of the obvious shortcomings of this approach:
- (1) time and effort to set up the needed infrastructure is high,
- (2) related work setups are created independently, leading to unnecessary duplication of work, but also heterogeneity, with certain details not documented impeding reproducibility.
The need to address such shortcomings has been identified for a long time, e.g., in the 2003 workshop of ACM SIGCOMM “Models, Methods and Tools for Reproducible Network Research” [1]. To overcome such challenges, the networked systems community built testbed RIs, such as GENI [2], FABRIC [3], Chameleon [4], CloudLab [5], and Fed4Fire [6].
While these initiatives demonstrated significant progress, many areas for improvement remain. Artifact evaluation committees need to spend significant effort to evaluate papers with artifacts (software and data with metadata) as experimental artifacts and the involved tools are highly customized and heterogeneous. At the same time, experimenters have to invest significant effort to prepare their artifacts for evaluation and reusability. This approach clearly does not scale, in particular in the context of data-driven science powered by AI/ML. An area for improvement is experiment control. So far, scientists who use a large testbed for specific experiments need to solve how to orchestrate their experiments, how to collect, process, and store the data produced by the experiment, and how to add metadata to support other scientists. Consequently, there is a high heterogeneity concerning the artifacts of specific experiments. The SLICES Research Infrastructure [7] introduced several concepts for the evolution of networking testbeds into scientific instruments, with comprehensive support for experiment control and data management.
The recent Dagstuhl seminar 24462 [8] entitled “Research Infrastructures and Tools for Collaborative Networked Systems Research” focused on bridging the gap between the services provided by large-scale testbed infrastructures and the needs of researchers conducting cutting-edge experiments. In this seminar, the networking community, the users of RIs, and the testbed community, the developers and providers of RIs and data management solutions, exchanged their viewpoints. The seminar concluded with a set of recommendations. A key goal of the non-paper session is to share and discuss these conclusions and recommendations with the SIGCOMM community. Conclusions to be discussed address strategic investment and community engagement, open access and data sharing, amplifying impact and network effects. Recommendations to be discussed include scientific objectives, common abstractions, reproducibility and repeatability, FAIR principles (Findability, Accessibility, Interoperability, and Reusability), flexibility and adaptability interoperability and openness, user experience, international collaboration and support of sustainable development goals (SDGs).
Planned agenda
- Short statements
- Georg Carle, Serge Fdida: Introduction to the Non-Paper-Session
- Serge Fdida, Georg Carle: Outcomes of Dagstuhl Seminar 24462
- Serge Fdida: Research instrument with experiment blueprints and data management
- Marie-Jose Montpetit: SLICES-RI: Research and Innovation
- Sebastian Gallenmüller, Damien Saucez: SIGCOMM and CoNEXT artifact evaluation and infrastructure needs
- Ethan Katz-Bassett: PEERING testbed
- Panel discussion
- Participants: Ellen Zegura and the above persons, including Q&A with the audience
References
- [1] Workshop on Models, Methods and Tools for Reproducible Network Research (MoMeTools), Program Co-Chairs: Georg Carle, Klaus Wehrle, Hartmut Ritter; Karlsruhe, Germany, August 25, 2003, In conjunction with ACM SIGCOMM 2003, https://conferences.sigcomm.org/sigcomm/2003/workshop/mometools
- [2] GENI. https://www.geni.net
- [3] FABRIC. https://fabric-testbed.net
- [4] Chameleon. https://www.chameleoncloud.org
- [5] CloudLab. https://www.cloudlab.us
- [6] Fed4Fire. https://www.fed4fire.eu
- [7] SLICES-RI. https://www.slices-ri.eu
- [8] Dagstuhl Seminar 24462 “Research Infrastructures and Tools for Collaborative Networked Systems Research”, https://www.dagstuhl.de/24462, report DOI: 10.4230/DagRep.14.11.60