A leading hub for innovative design and empirical understanding of scalable computing, data, and AI platforms that power software applications and emerging agentic ecosystems. We have expertise in operating systems, networks, databases, runtime environments, and distributed computing.
Our research spans across multidisciplinary research
This subtheme develops approaches of integrating AI and machine learning into the automatic monitoring and configuration of large-scale data management platforms. It includes automated benchmarking of data systems, self-driven query optimisation, automated physical data and index design, large-language models for query and schema generation, and agentic auto-tuning systems. The focus is on building scalable data platforms that can operate and adjust automatically to changing workloads and workload peaks.
This research strengthens the foundations of data processing platforms by improving the scalability, elasticity and fault tolerance of data platforms. It enables industry and government to build dependable data platforms for various use cases.
Professor Alan Fekete, Professor Albert Zomaya, Associate Professor Uwe Roehm, Dr Zhengyi Yang, Associate Professor Wei Bao, Associate Professor Kanchana Thilakarathna, Associate Professor Lijun Chang
This subtheme develops innovative architectures, algorithms and protocols for computing platforms including systematic approaches to testing, validating, and securing software platforms, from embedded and IoT systems to agentic and AI-powered environments. It includes grammar-based fuzzing for automated input generation, mutation analysis for evaluating test effectiveness, statistical methods for coverage estimation, and security testing of industrial control system protocols. The focus is on building rigorous, evidence-based methods that ensure computing platforms behave correctly and securely under diverse and adversarial conditions.
This research strengthens the dependability and security of software platforms across critical domains including industrial control systems, IoT infrastructure, and emerging agentic ecosystems. It provides industry and government with principled tools and techniques for systematic vulnerability detection and test quality measurement, enabling more trustworthy deployment of software-intensive platforms.
Dr Hong Jin Kang, Dr Rahul Gopinath, Dr Xi Wu, Dr Jiangshan Yu, Associate Professor Qiang Tang, Associate Professor Kanchana Thilakarathna
This subtheme develops methods for improving how software agents can be designed, implemented, tested, evaluated, and maintained. Agentic systems combine large language models with prompts, tools, harness and context management methods, and their behavior is difficult to understand and debug using traditional software engineering methods. This focus is on establishing methods and tools for engineering reliable agentic systems.
This research strengthens the reliability and security of agentic platforms and their ecosystem, including models, agentic frameworks, and the supply chain of AI agents. It provides industry and government with techniques for assessing the reliability and building agentic systems.
Dr Hong Jin Kang, Dr Rahul Gopinath, Dr Xi Wu, Dr Jiangshan Yu, Associate Professor Lijun Chang, Dr Zhengyi Yang
We develop novel benchmarks, with innovative workloads and metrics, to capture more aspects of system performance than existing measurements. We also explore ways of improving how one can automatically run benchmarks, and ways to automatically generate datasets and workloads for benchmarks.
This research helps researchers and practitioners make sound decisions about system choices, which more accurately reflect the real concerns that drive application performance.
Associate Professor Uwe Roehm, Professor Alan Fekete, Professor Vincent Gramoli, Dr Jiangshan Yu
This subtheme develops efficient and scalable algorithms and systems for analysing large-scale graph data, which naturally model interconnected relationships between entities. It includes graph mining, graph querying, graph indexing, and scalable methods for analysing sparse, dynamic, and complex graphs. The focus is on developing high-performance graph analytics techniques and platforms that enable efficient analysis of large and evolving networked datasets.
This research strengthens the scalability and efficiency of graph data analytics, enabling the analysis of large and complex networked data at scale. It provides industry and government with advanced methods for uncovering hidden structures, analysing connectivity and communities, and supporting data-driven decision making across domains such as cybersecurity, infrastructure, transport, and social networks.
Associate Professor Lijun Chang, Dr Zhengyi Yang
