Thought experiments, data and reproducibility for networking and FutureG research: Hackathon

    WORKSHOP/HACKATHON DESCRIPTION

    Digital infrastructures are key drivers of societal digital transformation. Research and innovation in this domain address critical industrial demands, sovereignty and security challenges, and opportunities for economic growth and societal advancement.

    Research in digital science, particularly in networking, is deeply intertwined with complex systems in which both software and hardware play critical roles. Consequently, advancing experimental research in this field requires conducting experiments on real-world systems.

    However, access to such systems remains a significant challenge for many researchers due to the substantial costs and technical barriers involved. Specialized equipment, such as advanced radios, smart NICs, or GPUs, are prohibitively expensive. In addition, operating and managing these systems often require highly specialized technical expertise, further limiting accessibility and participation in experimental research.

    The SLICES Research Infrastructure (SLICES-RI) [1] is designed to facilitate experimental research. It offers long-term access to large-scale, distributed, state-of-the-art hardware and supports research activities. SLICES-RI uses blueprints and their associated support to allow researchers to minimize the effort required to replicate and conduct complex experiments, according to the work carried out by the community [2]. The blueprints provide reproducible tools, software, hardware, and methodologies designed to support robust experimental research using state-of-the-art infrastructures. This allows researchers around the world to concentrate on their core research while benefiting from the infrastructure made available by the community

    SLICES-RI supports the entire experiment lifecycle by deploying appropriate tools to facilitate user-friendly and robust experimentation. As shown in Fig. 1, experiments in SLICESRI are orchestrated by the plain orchestrating service (pos) [3], which manages the experiment, its setup, and evaluation. This includes the recording of all data and meta-data generated by the experiment to allow sound analysis and reproducibility. This workflow is used in the post-5G blueprint, which deploys a fully functional 5G network with a cloud native core [4] and a split 7.2 radio access network [5]. The blueprint is flexible enough to allow its deployment in fully virtual environments (e.g., in a public cloud) or with dedicated physical hardware on SLICES premises.

    In this hackathon, organized for the second year, we plan to bootstrap new software and features for the research community.

    On the one hand, we will utilize packaging techniques to simplify experiments for students and researchers who want to run 5G and beyond 5G experiments. For example, the audience will be provided with instructions on how to code an experiment using Ansible, to control the experimentation process (e.g., module for pos [6]), thus removing the need to learn low-level experiment syntax, and reducing the learning curve. Another contribution could be to automate the provisioning of cloud infrastructure with Terraform. In its current form, Terraform recipes are written for several types of resources in a non-uniform manner. A unified approach would allow easy migration from one type of infrastructure (e.g., using a workstation) to another (e.g., a complex distributed environment with on-site and off-premises resources).

    On the other hand, we would like to contribute to the opensource software community by adding features that would be beneficial for SLICES-RI and the blueprint. For example, we could add smooth support for OpenID in OpenVPN and link it to RBAC in kubernetes clusters, for example, with the help of Keycloak. Another example would be to improve the OpenAirInterface platform documentation for the CU/DU separation. Another example would be to create a data model to automatically generate OpenAirInterface 5G configurations. Integration of new hardware (e.g. support for new RF-boards for the radio) is also welcome.

    In summary, the objective of the hackathon is to improve the implementations of different blueprints for SLICES, and at the same time be beneficial for the entire research community beyond SLICES. Moreover, the automation of experiments and the clear definition of blueprints within SLICES-RI present a significant opportunity to establish Trustworthy Reproducibility. While conducting experiments in a reproducible framework ensures that procedures can be repeated, it does not inherently guarantee the validity of the results. This is where the concept of trustworthy reproducibility becomes essential. In this year’s hackathon, we propose to initiate the development of a foundational framework for trustworthy reproducibility. Important notice: This hackathon is a companion activity of the Data-Driven Research, Reproducibility, and FAIR Practices in Future G Networks workshop.

    The workshop and hackathon are designed to complement each other, creating a synergistic dynamic. The workshop focuses on test platforms at a broad level, while the hackathon aims to bootstrap innovative collaborative tools for these platforms.

    The two events target distinct but interconnected audiences and nicely complement each other. The workshop is expected to attract a more traditional audience, including researchers, decision-makers, and post-docs. In contrast, the hackathon, by its hands-on nature, will primarily draw research engineers and developers—professionals who often collaborate closely with the workshop’s participants. This complementary setup fosters a fruitful exchange of ideas and expertise across both events.

    PLANNED FORMAT OF THE WORKSHOP

    The workshop will be conducted interactively, actively engaging the audience with different activities. It comprises five main parts:

    • A first session of 30 minutes will introduce SLICES and the blueprints from a technical standpoint;
    • A 30-minute session will then present the concepts of Trustworthy Reproducibility;
    • A 2h session will then be used to define the projects to work on during the rest of the day and set the teams. During this session each team will present their proposal in a lightning talk of 180 seconds;
    • The rest of the day will be dedicated for code sprinting on the projects;
    • In the evening, each team will make a 5-minute demo of their project and the committee will elect the winner of the hackathon

    As indicated above, the program aims at lively and interactive discussions with the participants and stakeholders involved. Teams will be supervised by experienced engineers specialized in agile methodology, incorporating code sprints in their day-to-day practice. Project proposals are to be submitted by email prior to the workshop with the submission of an Expression of Interest (EoI) to the workshop co-chairs in the form of an abstract that must explain:

    • a proposed team of a maximum of 4 persons (minimum 1 person) with the identification of its leader;
    • a 200-word summary of the project;
    • a work plan with the different milestones to be achieved during the day of the hackathon and potential extensions;
    • an explanation of why the project would be beneficial to SLICES-RI and to the whole research community.
    Descriptive text of the image

    We plan for teams to be composed of up to four participants. The exact synthesis of the teams will be decided on the morning of the event, based on the hackathon projects decided collectively.

    Due to funding constraints, we focus on providing valuable resources, including coffee breaks, meals and networking opportunities.

    PAST EVENTS

    We successfully organized the first workshop in IFIP Networking last year, where 5 teams worked on diverse subjects, such as the integration of CEPH distributed storage to the post5G blueprint, support of RBAC in the infrastructure, or bug correction.

    SLICES-RI is a long-term project (2021-42) and this hackathon has the opportunity to be the second dedicated to the project. Having already organized two successful events (the first and second SLICES Summer Schools in 2022 and 2023) make the committee confident that the hackathon will be of interest to the research community. In particular, in 2022 we organized the first SLICES Summer School (https://www.slices-ri.eu/events/slices-sc-summer-school/) that gathered about 60 attendees moving between tutorials and hands-on. The second SLICES Summer School took place in 2023 (https://www.slices-ri.eu/events/slices-sc-2nd-summer-school/) and gathered 80+ attendees over 3 days with more time dedicated to hands-on. It is now a recurrent flagship event for SLICES-RI and the third summer school was organized in summer 2024 with 70 participants over a duration of 5 days.

    In parallel to physical events, the SLICES-Academy [7] offers a detailed tutorial on deploying the blueprint. About 40 students follow this class and a dozen regularly practice the class.

    SUBMISSION GUIDELINES

    Project proposals are to be submitted by email with the submission of an Expression of Interest (EoI) to the workshop co-chairs (This email address is being protected from spambots. You need JavaScript enabled to view it., This email address is being protected from spambots. You need JavaScript enabled to view it., and This email address is being protected from spambots. You need JavaScript enabled to view it.) in the form of an abstract that must explain:

    • a proposed team of a maximum of 4 persons (minimum 1 person) with the identification of its leader;
    • a 200 words summary of the project;
    • a workplan with the different milestones to be achieved the day of the hackathon and potential extensions;
    • an explanation of why the project would be beneficial to SLICES-RI and to the whole research community.

    The format can be a PDF or a link to a git repository, clearly documented. The name and email address of each team member must be provided in the EoI.

    To promote openness and reproducibility authors and reviewers are not anonymous and it is expected to have interactions between the authors and reviewers during the review process.

    All code produced during the hackathon will be done under a public, free, and open source license.

    Team members engage to physically attend the hackathon if their application is selected.

    IMPORTANT DATES

    • Expression of Interest (EoI) submission: 21 March 2025.
    • Feedback on EoI (notification): 25 April 2025.
    • Final Lightning talk slides: 2 May 2025.
    • Hackathon date: XX June 2025.

    TCP CO-CHAIRS

    1. Panayiotis Andreou, UCLan Cyprus, (This email address is being protected from spambots. You need JavaScript enabled to view it.)
    2. Nikos Makris, University of Thessaly, (This email address is being protected from spambots. You need JavaScript enabled to view it.)
    3. Damien Saucez, Inria, (This email address is being protected from spambots. You need JavaScript enabled to view it.)

    TECHNICAL PROGRAM COMMITTEE

    Find below an initial TPC proposal (members with a ∗ are confirmed).

    • Armir Bujari (University of Bologna)
    • Attila Farkas (SZTAKI)
    • Serge Fdida (Sorbonne Universite)*
    • Sebastian Gallenmuller (TUM)*
    • Carmen Guerrero (UC3M)
    • Nikos Makris (UTH)*
    • Raymond Knopp (Eurecom)*
    • Andrea Passarella (IIT)*
    • Michele Polese (Northeastern University)
    • Esa Posio (Oulu University)*
    • Damien Saucez (Inria)*
    • Ivan Seskar (Rutgers University)
    • Thanasis Korakis (UTH)*
    • Brecht Vermeulen (IMEC)

    REFERENCES

    [1] “SLICES-RI — Scientific LargeScale Infrastructure for Computing/Communication Experimental Studies — slices-ri.eu.” https://www.slices-ri.eu/. [Accessed 08-01-2024].

    [2] SLICES, “SLICES-RI blueprint.” http://doc.slices-sc.eu/blueprint/, 2024.[Accessed 22-01-2024].

    [3] S. Gallenmuller, D. Scholz, H. Stubbe, and G. Carle, “The pos framework: ¨A methodology and toolchain for reproducible network experiments,” in Proceedings of the 17th International Conference on Emerging Networking EXperiments and Technologies, CoNEXT ’21, (New York, NY, USA), p. 259–266, ACM, 2021.

    [4] N. Nikaein, M. K. Marina, S. Manickam, A. Dawson, R. Knopp, and C. Bonnet, “OpenAirInterface: A Flexible Platform for 5G Research,”SIGCOMM Comput. Commun. Rev., vol. 44, p. 33–38, oct 2014.

    [5] V. Q. Rodriguez, F. Guillemin, A. Ferrieux, and L. Thomas, “Cloud-ran functional split for an efficient fronthaul network,” in 2020 International Wireless Communications and Mobile Computing (IWCMC), pp. 245–250, 2020.

    [6] S. Gallenmuller, D. Scholz, H. Stubbe, and G. Carle, “The pos framework: ¨a methodology and toolchain for reproducible network experiments,” in Proceedings of the 17th International Conference on Emerging Networking EXperiments and Technologies, CoNEXT ’21, (New York, NY, USA), p. 259–266, Association for Computing Machinery, 2021.

    [7] SLICES, “Slices Academy — Scientific LargeScale Infrastructure for Computing/Communication Experimental Studies — slices-ri.eu.” https://www.slices-ri.eu/slices-academy/, 2024. [Accessed 08-01-2024].