Periodic Reporting for period 4 - SyNET (From Network Verification to Synthesis: Breaking New Ground in Network Automation)
Reporting period: 2024-05-01 to 2025-10-31
The project is addressing the general challenge of building provably-correct network infrastructures. This is an important challenge as, each year, billions of Euros are lost due to avoidable, human-induced network downtimes. Worse yet, some of these downtimes affect critical infrastructures such as the one supporting emergency calls. The unique approach of this project is to leverage formal network models to automatically synthesize correct and efficient network configurations from high-level requirements. In addition to synthesizing correct configurations, these models will also be used to verify running network infrastructures, thereby allowing to improve existing network deployments. Technically, the project sits at the intersection of several areas of computer sciences including: computer networks, programming languages, and formal methods. Regarding concrete objectives, the project aims at developing an end-to-end configuration synthesizer platform which is: (i) optimal (it generates provably-optimal configurations); (ii) easy-to-use (for network operators); and (iii) expressive (it supports a vast collection of network protocols and mechanisms).
Over the lifetime of the project, we realized the vision of moving from network verification to synthesis, delivering methods and systems that generate and validate network configurations with strong guarantees on correctness, robustness, and performance. Concretely, we developed: (i) synthesis techniques for safe network-wide configuration updates (SIGCOMM) and for correctness-preserving BGP reconfiguration under transient dynamics (SIGCOMM); (ii) methodologies to increase trust in the verification toolchain by systematically uncovering modeling errors (NSDI); (iii) foundational results clarifying what makes network-wide configuration synthesis hard or tractable (ICNP); (iv) techniques to detect and explain transient forwarding anomalies and to predict convergence-time violations (CoNEXT); and (v) the first practical verification of worst-case link-load properties (NSDI). These works were complemented by further works on scalable exploration, causality analysis, and probabilistic reasoning (ICNP), as well as on generating representative traffic for evaluation (HotNets).
Dissemination and exploitation were achieved through 11 peer-reviewed publications and presentations in top-tier venues in computer networks and systems (including SIGCOMM, NSDI, CoNEXT, and ICNP), and through practical artefacts and prototypes (all released in open-source), enabling reproduction and providing a basis for technology transfer. The results also led to interactions with stakeholders (including network operators) and contributed to the training and graduation of multiple PhD students, with several works receiving major recognition and awards.
Dissemination and exploitation were achieved through 11 peer-reviewed publications and presentations in top-tier venues in computer networks and systems (including SIGCOMM, NSDI, CoNEXT, and ICNP), and through practical artefacts and prototypes (all released in open-source), enabling reproduction and providing a basis for technology transfer. The results also led to interactions with stakeholders (including network operators) and contributed to the training and graduation of multiple PhD students, with several works receiving major recognition and awards.
SyNET pushed beyond the state of the art by turning network configuration management into a correct-by-construction process. Traditionally, operators rely on error-prone, best-effort (often manual) procedures, and human mistakes cause a large fraction of network outages; SyNET showed that it is possible to automatically compute network-wide configurations that provably guarantee important properties, even for complex routing protocols.
SyNET advanced the state of the art in multiple complementary ways. First, SyNET enabled correctness over time, with guarantees not only after convergence but also during transients while the network adapts to failures and changes—when outages often occur. Second, it advanced performance reasoning with guarantees by predicting and bounding worst-case congestion and traffic hotspots under failures and routing changes. Third, it strengthened the trustworthiness of automation by systematically uncovering (and helping fix) errors in network verification models.
Overall, SyNET leaves validated methods, prototypes, and foundations that support more reliable (“always-on”) connectivity and move network operations toward automation that is auditable and safe.
SyNET advanced the state of the art in multiple complementary ways. First, SyNET enabled correctness over time, with guarantees not only after convergence but also during transients while the network adapts to failures and changes—when outages often occur. Second, it advanced performance reasoning with guarantees by predicting and bounding worst-case congestion and traffic hotspots under failures and routing changes. Third, it strengthened the trustworthiness of automation by systematically uncovering (and helping fix) errors in network verification models.
Overall, SyNET leaves validated methods, prototypes, and foundations that support more reliable (“always-on”) connectivity and move network operations toward automation that is auditable and safe.