The Network and Systems (NS) unit conducts cutting-edge research on the design, analysis, and operation of modern distributed systems. Our work spans the full technology spectrum, from resource-constrained IoT devices and safety-critical Cyber-Physical Systems (CPS) to high-performance cloud platforms and decentralized online services. We develop foundations and systems that enable secure, efficient, and trustworthy computation, communication, and data processing at scale.
Research Themes
Our research builds on strong expertise in distributed algorithms, systems security, stream processing, machine learning systems, and resource management. Our research advances several interconnected areas, reflected in recent work published at leading international venues:
Reliable and Efficient Distributed and Parallel Algorithms
We advance the theory and practice of resilient distributed computation, including self-stabilizing and Byzantine-tolerant protocols, fault-tolerant broadcast and consensus, efficient concurrent data structures to facilitate distributed, parallel and and stream processing with possibilities to tune trade-offs between guarantees and resource requirements, and parallel stochastic optimization. Our work enables systems to operate correctly despite faults, asynchrony, or adversarial conditions.
Security and Privacy in Connected Systems
We develop techniques to secure modern digital ecosystems, from vehicular networks and industrial IoT to blockchains, federated learning, and edge AI. Recent results include frameworks for secure remote attestation, privacy-preserving large-scale monitoring, attack detection in in-vehicle networks, secure coding evaluation, and network-level privacy attacks and defenses. We also investigate cryptocurrency security, blockchain consensus robustness, and censorship or hijack attacks in large-scale networked systems.
High-Rate Stream Processing and Online Analytics
Our work advances the state of the art in elastic, fault-tolerant stream processing, sketch-based data structures, real-time vehicular analytics, and provenance tracking. We design algorithms and frameworks (with open-source code) for efficient monitoring of IoT- and CPS-scale data, enabling real-time insights even under extreme data rates and distributed deployments.
Machine Learning and Operating Systems from the Edge to the Cloud
We investigate the performance, sustainability, and robustness of ML systems, including reducing ML system bloat, secure ML pipelines, adaptive model selection for live video analytics, and foundation models for time-series forecasting. Our contributions include edge-native model serving, privacy-preserving inference, and high-performance ML-powered traffic classification.
Data Management and Data-Driven Support in Cyber-Physical Systems
We develop data-driven methods for the efficient and sustainable operation of cyber-physical systems (CPS) and large-scale distributed infrastructures. Our work spans energy systems, vehicular platforms, IoT environments, and cloud infrastructures, integrating resource management, optimization, and system-level intelligence. Key topics include elastic service scaling, carbon-aware scheduling, power-aware cloud workload management, and peer-to-peer energy-sharing optimization, supported by measurement-driven modeling and real-time control.
Approach and methodology
Our research spans rigorous algorithmic design and theoretical analysis to experimental systems research. We combine principled modeling with hands-on implementation, conducting large-scale deployment studies, measurement-driven modeling, and real-world evaluations across vehicular systems, IoT platforms, and cloud infrastructures.
A central aspect of our methodology is building and operating real systems. Many of our prototypes and research artifacts are developed within collaborative projects and are released as open-source software whenever possible. We view open sourcing not only as a means of dissemination, but as a way to enable reproducibility, foster community engagement, and accelerate technology transfer into practice. Through this integrated approach, from theory to deployed systems, we ensure that our research advances are both scientifically rigorous and practically relevant.
Collaboration and impact
The NS unit maintains long-standing collaborations with leading industrial partners such as Ericsson, Volvo AB, Volvo Cars, and Göteborg Energi, as well as top universities worldwide. Many of our research results are deployed in operational systems, influencing industry standards and practices. The unit is actively involved in major projects funded by the European Commission, the Swedish Research Council (VR), the Wallenberg Foundation, the Wallenberg Autonomous Systems Program, Vinnova, and the Swedish Foundation for Strategic Research (SSF).
Teaching and education
We contribute to core curriculum courses on the Bachelor level for the Computer Science and Engineering programme. We are also heavily engaged in advanced education within the Computer Systems and Cybersecurity master’s programme, covering modern operating systems, distributed systems, computer networks, data communication, network and computer security, and data-driven support in cyberphysical systems (communication, energy, production, vehicle/transport-systems).
Our research directly informs course content, ensuring students gain hands-on exposure to state-of-the-art methods, tools, and systems. Every year we offer projects on both the Bachelor and Master level.
