2nd ACM SIGCOMM Workshop on Future of Internet Routing & Addressing (FIRA)
Sunday, September 10, 2023
9:15am-11:00am Session 1 on “Internet Technologies”
Keynote on “Now We Understand How the Internet Is Evolving--What's Next?”
Speaker: Pamela Zave and Jennifer Rexford (Princeton University)
Abstract: This community shares an understanding of how the Internet has evolved, since its creation, to meet challenges far beyond its original goals. The understanding is embodied in the CFP's reference to "limited domains,” and in our new model Compositional Network Architecture (presented in our forthcoming book). In this talk we explain how the new model provides a rigorous, modular foundation for these ideas. We then discuss a range of research questions, building on what has been learned so far.
Bio: Pamela Zave is a researcher with the Computer Science Department of Princeton University, having held previous positions at the University of Maryland, Bell Labs, and AT&T Labs. She is an ACM Fellow, an IFIP Fellow, an AT&T Fellow, and the 2017 recipient of the IEEE Harlan D. Mills Award for sustained contributions to the theory and practice of software engineering. She works on the application of formal methods to networking, and is an author of the forthcoming book ‘The Real Internet Architecture: Past, Present, and Future Evolution.
Homomorphic Routing: Private Data Forwarding in the Internet
Francesco Tusa, David Griffin and Miguel Rio
Peer Clustering for the InterPlanetary File System
Y. Thomas, N. Fotiou, I. Pittaras, G. Xylomenos, S. Voulgaris, G. C. Polyzos
11:15am-1:00pm Session 2 on “Challenges in New Networking Environments”
Keynote on “The SCION Research and Education Network”
Speaker: Adrian Perrig (ETH Zurich)
Abstract: Deployment of a new internet architecture appears almost infeasible, faced by a long road to wide-spread deployment. A large part of the challenge are dependency loops among a number of constituents: end users, leaf (autonomous system) domains, ISPs, OS writers, and application developers. End users rely on their leaf domain and applications or OS supporting the new architecture -- but their adoption is driven by end user demand. Similarly, ISPs require leaf domain demand, but leaf domain demand can only emerge once ISPs offer the new technology. To break these dependency cycles, we suggest establishing 1'000'000 end hosts with access to the new network architecture -- in the anticipation to provide sufficient incentives for a virtuous cycle of adoption to get started. In this talk, we will discuss how we plan to achieve this with the SCION research and education network.
Bio: Adrian Perrig is a Professor at the Department of Computer Science at ETH Zürich, Switzerland, where he leads the network security group. He is also a Distinguished Fellow at CyLab, and an Adjunct Professor of Electrical and Computer Engineering at Carnegie Mellon University. From 2002 to 2012, he was a Professor of Electrical and Computer Engineering, Engineering and Public Policy, and Computer Science (courtesy) at Carnegie Mellon University. From 2007 to 2012, he served as the technical director for Carnegie Mellon's Cybersecurity Laboratory (CyLab). He earned his MS and PhD degrees in Computer Science from Carnegie Mellon University, and spent three years during his PhD at the University of California at Berkeley. He received his BSc degree in Computer Engineering from EPFL. He is a recipient of the ACM SIGSAC Outstanding Innovation Award. Adrian is an ACM and IEEE Fellow. Adrian's research revolves around building secure systems --in particular his group is working on the SCION secure Internet architecture.
Performance of Meshed Tree Protocol in Data Center Networks
Peter Willis, Miaoxin Li, Nirmala Shenoy
Seamless Hardware-Accelerated Kubernetes Networking
Mircea M. Iordache-Sica, Tula Kraiser, Olufemi Komolafe
2:00pm-3:45pm Session 3 on “Traffic Engineering”
Keynote on “"Navigating Network Telemetry: Efficiency, Insights, and Routing Challenges”
Speaker: Ran Ben Basat (UCL)
Abstract: With recent progress in programmable devices, we can now capture fine-grained network telemetry, which serves diverse functions like congestion control and identifying flow performance problems. Nonetheless, the extensive volume of information such telemetry generates leads operators to employ it very selectively. This talk will explore the emerging methods for enhancing telemetry, including coding the information to reduce the per-packet overheads, using CPU-bypass techniques to scale collection, automatically identifying flows with poor performance to trigger telemetry collection, and routing telemetry away from congestion hotspots to avoid interfering with user traffic.
Bio: Ran Ben Basat is a Lecturer (Assistant Professor) at University College London's Computer Science Department. He holds a Ph.D. from Technion and was a postdoctoral fellow at Harvard University. His expertise lies in algorithms for networking and machine learning and he has been recognized with the Meta Network for AI Faculty award.
If Iterative Diffusion Is The Answer, What Was The Question?
David Guzman, Dirk Trossen, Joerg Ott
QCMP: Load Balancing via In-Network Reinforcement Learning
Changgang Zheng, Benjamin Rienecker, Noa Zilberman
4:00pm-4:45pm Panel on “Better traffic engineering for interconnected networks” with keynote speakers
Moderator: Dirk Trossen
4:45pm-5:00pm Closing words
Call for Papers
Applications are placing increasingly sophisticated demands on the network for better quality, more predictability, and greater reliability. Some of these applications are futuristic predictions (for example, holographic conferencing, immersive digital worlds, intelligent industry, telemedicine, automated agriculture, and sensory Internet), while existing applications are already seeing real network demands (such as multiplayer augmented- or virtual- reality games, industrial networking, and distributed AI). This overlaps with a growing trend to extend end-to-end communications to include highly computational processes (e.g., emergent intelligent application), machines and moving objects (such as autonomous cars and drones), while increasingly utilizing the ability to virtualize and replicate any network (digital twin) or service for improved efficiency and redundancy. Also, new environments, such as (beyond) 5G, manufacturing, or Non-Terrestrial Networks (NTN) pose new challenges to the requirements, such as time-variance of connectivity availability, At the same time, integrating connectivity with computational services is seen as another important consideration to improve the overall system’s efficiency and, ultimately, carbon footprint. This has led to many investigations to extend the most basic packet delivery mechanism that the Internet currently provides, reflected in both routing and addressing behaviors, often specific to a particular use case and thus addressing a specific set of application traffic requirements or networking characteristics.
Many proposals have been made to go beyond basic locator-based routing, often adding information to IP packets or even entirely replacing the current IP packet delivery architecture with a new one (which may or may not include IP-like packet delivery). The intent is always to facilitate enhanced routing and forwarding decisions to provide differentiated behaviors for different packet flows distinct from simple shortest path first routing in basic IP packet delivery. Moreover, dedicated networking environments, such as data-center networks, space networks, vehicular networks, CDNs and others have driven the creation of a plethora of solutions, each of which improves routing and addressing within the requirements and specific behaviors of those environments. RFC8799 captures this phenomenon as proliferation of limited domains, interconnected via the public Internet as we know it.
Beyond the evolving purpose of communication systems, in the form of novel applications and novel network methods to accommodate them, a growing direction of work investigates the use of AI-based learning methods for finding ‘best’ (or better) ways to communicate, enhancing or even (partially) replacing established methods of protocol engineering and system operation.
Last year’s inaugural FIRA workshop showed firsthand the interest that the research community has in discussing their ideas and proposals for evolving the basic packet delivery capability of the Internet. A strong program of research works not only showed that routing and addressing is a key topic of research, but these discussions were directly bridged to engineering communities in the IETF and IRTF through side meetings and contributions, thus closing the gap between cutting edge research and opportunity for future engineering development that will lead to deployment.
Beyond the continued interest in the R&D community, public funding has increasingly focused on the development of new network solutions, particularly in the context of the accelerating developments for 6G. The successful start of the EU research & innovation framework programme "Horizon Europe" and its Smart Networks and Services (SNS) Joint Understanding, has seen first efforts in publicly funded network research being established, while upcoming funding rounds continue to call for work on 'improving data plane performance'. In addition, industry interest has continued to increase, within initial efforts such as the 6G Networking whitepaper and the inaugural 6G Networking Symposium in Lisbon, leading to the formation of the 6G Industry Association in Europe and counterparts in other regions of the world. The FIRA workshop will continue to build on this industry- and academia-spanning interest in novel network solutions. However, this interest also poses a problem in that those many solutions have often been developed in isolation and without a larger and coherent architectural view of evolving the Internet as a whole.
This year’s Future of Internet Routing and Addressing (FIRA) workshop aims to build on the success of last year’s event by soliciting experiences and insights in developing the requirements-driven architectural proposals that were previously discussed, and highlighting research topics and findings that support proposals for incremental or radical changes to the Internet routing and addressing paradigms. In other words, this second FIRA workshop is looking to investigate and expand the boundaries of what can be achieved by introducing new architectures rather than point solutions, those architectures extending existing or inventing new routing and addressing techniques that modify the default forwarding behavior to be based on other information present in the packet, said behavior being either configured policy or dynamically programmed into the routers and devices. These new forwarding behaviors aim at causing new and alternative path processing by routers that aim at, among others:
- Determinism of quality of delivery in terms of throughput, latency, jitter, drop precedence
- Support for resilience in terms of survival of network failures and delivery degradation
- Support for highly distributed, virtualized service execution environments, where route changes are aligned (in time) with the ability to establish new service execution points
- Support for new or enhanced addressing schemes that may require enhanced or new forwarding behaviours
- Support for multi-constraint and service- and/or compute-aware routing as well as forwarding behavior
- Support for programmability of forwarding behavior as an alternative to SW virtualization of Layer 3 network functions
- Energy efficiency considerations as a key performance indicator for new forwarding behavior
- Time-variable routing and forwarding decisions in response to planned loss, or activation, of network connectivity based on scheduled events
- Consideration for improving on user and system security (including privacy)
- Using AI-based methods to learn best or better forwarding behaviours in advancement of a priori determined engineering solutions
- Improvement of routing performance in terms of the volume of data that has to be exchanged both to establish and to maintain the routing tables
- Deployability in terms of configuration, training, development of new hardware/software, and interaction with pre-existing network technologies and uses
- Enabling zero-touch mechanisms, e.g., autonomous control plane approaches, for routing
- Efficiency of manageability in terms of, i. diagnostic management, ii. management of Service KPIs with/without guarantees, and iii. dynamic and controlled instantiation of management information in the packets
- Monetization of network resources, particularly in the light of delivering advanced applications to end users
The FIRA workshop solicits work on use cases, design principles, architectures, and techniques, but also solicits insights into implementations and experiments with novel solutions that address the highlighted objectives.
The workshop will start with a keynote address to scope the research agenda. This will lead to topical, cornerstone keynotes, followed by a session of paper presentations to provide updates on state-of-the-art research and allow questions and debate. Finally, the workshop will conclude by combining the topics into a panel session.
Topics of Interest
- Advanced applications and use case analysis and requirements
- Goals and challenges in future and evolving routing and addressing schemes
- Architecture frameworks for multi-purpose routing
- Limited domain architectures
- Limited domain interconnection architectures
- Routing on multiple optimality criteria
- Compute-aware routing and forwarding
- Low latency forwarding techniques
- Time-variable routing and forwarding
- Enhanced host protocol stacks for advanced routing and forwarding solutions
- Routing on new forms of identification
- Coordination of information and decisions across multiple domains (regions & technology layers)
- Time-variant routing mechanisms and solutions
- Routing based on formal routing algebras and regular expressions approaches
- Routing for constrained and intermittently connected environments
- Routing and forwarding based on AI-based methods, such as federated learning
- Routing and forwarding based on zero-touch configuration
- Centralized routing architectures
- Integration with modern SDN architectures and protocols
- Programmable forwarding architectures
- Security analysis of semantic enhancements
- Scalability analysis of novel routing mechanisms and semantic enhancements
- Impact of semantic enhancements on privacy
- Economic and game-theoretical analysis of enhanced network semantics
- Analysis on impact of enhanced routing semantics on net neutrality
- Commercial and strategic cost/benefit analyses
- Information and data models for adoption
- Stability design and analysis
- Experience and deployment
Submissions must be original, unpublished work, and not under consideration at another conference or journal. Submitted papers must be at most six (6) pages long, including all figures, tables, references, and appendices in two-column 10pt ACM format.
We also encourage research demos for which a two-page extended abstract must be submitted in the same format as the workshop papers.
Papers and extended abstracts must include author names and affiliations for single-blind peer reviewing by the PC. Authors of accepted submissions are expected to present and discuss their work at the workshop.
Please submit your paper via https://fira2023.hotcrp.com/.
If you have any questions or problems with your submission, please get in touch with Dr. Dirk Trossen (firstname.lastname@example.org).
June 11, 2023
July 2, 2023
July 16, 2023
- Steering Committee (SC)
University of Cambridge
Old Dog Consulting
- Technical Program Committee (TPC) Chairs
University College London
- Technical Program Committee (TPC) Members
Huawei Research Labs
University of Napoli
University College London
Tokyo City University
Centre Tecnològic Telecomunicacions Catalunya
Huawei Research Labs