Research Areas

Our research focuses on advancing computing systems by enhancing performance, availability, reliability, cost-efficiency, power consumption, and runtime operations. Our approach combines mathematical modelling with large-scale cloud experimentation, and we frequently develop proofs-of-concept and prototypes.
🖥️ Serverless Computing

Performance modelling, simulation, and optimization of serverless (FaaS) platforms — including cold-start analysis, autoscaling, function placement, and cost/performance trade-offs.

🤖 Machine Learning Systems

Predicting deep neural network training time, convergence prediction across architectures, SLA-aware inference serving, and ML-based autoscaling for cloud applications.

🧩 Microservice Platforms

Performance modelling of microservice platforms, software diversity and multi-versioning to engineer reliable and performant systems, and efficient provisioning of microservices.

🔗 Cyber-Physical Systems

Self-managing IoT platforms, end-to-end IoT application management, intrusion detection for smart grids, and big-data analytics for smart transportation.

☁️ Cloud Software Systems

Adaptive and self-managing cloud applications, elastic containerized applications for DevOps, autoscaling and monitoring as a service, and analytical performance/availability models of cloud data centers.

⚙️ Operating Systems

Runtime operations, resource management, and controller-based optimization (e.g., adaptive PID controllers) for containerized software systems.

⛓️ Distributed Ledgers

Performance evaluation of blockchain systems, DAG-based distributed ledgers for smart communities, blockchain-based serverless platforms, and peer-to-peer energy trading frameworks.

Related Publications

(2026). ORION: Integrated Runtime Modelling for Predicting Deep Learning Training Time. In Proceedings of the 17th ACM/SPEC International Conference on Performance Engineering, Florence, IT. ACM.

(2025). PreNeT: Leveraging Computational Features to Predict Deep Neural Network Training Time. In Proceedings of the 16th ACM/SPEC International Conference on Performance Engineering, pages 81-91, Toronto ON Canada. ACM.

(2024). A Learning-Based Caching Mechanism for Edge Content Delivery. In International Conference on Performance Engineering (ICPE ‘24). ACM.

(2024). Employing Software Diversity in Cloud Microservices to Engineer Reliable and Performant Systems. In IEEE International Conference on Autonomic Computing and Self-Organizing Systems (ACSOS ‘24). IEEE.