I work on data-driven approaches to improve the performance and programmability of distributed systems. In the context of cloud computing, this entails looking into API interoperability and migration technologies, as well as related decision support issues. These and more challenges are collectively referred to as cross-cloud computing. This also extends to post-deployment composition of systems of systems.
I am interested in fog computing, particularly adaptive edge resource placement using micro-clouds. I am equally interested in resource-rich deployments (e.g. IoT) and resource-poor ones (e.g. data centre-remote locations).
I also work on border-free network architectures in intent-driven systems. Finally, I advocate network-awareness which involves measuring networked systems, evaluating network protocols, and proposing new network management strategies.
Current projects
-
Address the key challenges for personalised media experiences, particularly the creation and delivery at scale, using AI and Object-Based Media (OBM).
Funder:
EPSRC
Value:
£3.1m
Key outputs:
• User customization options in media can both improve and harm viewer satisfaction depending on how they're implemented and used.
• Changing video playback speed affects viewer quality of experience differently than streaming quality degradation.
• Viewers experience different levels of satisfaction when watching picture-in-picture gaming content depending on which video element receives focus.
• Evaluating viewer satisfaction with OBM (media that combines multiple video objects) requires different assessment methods than traditional single-video approaches.
Past projects
ABC: 1, 2
-
Selecting cloud services through the introduction of a smart and continuously adaptive cloud broker.
Funder:
EPSRC
Value:
£775k
Key outputs:
- Using transfer learning to reduce the cost of predicting the performance of different cloud applications on various cloud infrastructures.
- Providing concepts for modelling service level objectives of cloud applications.
- A domain-specific modeling language to support the evolution of multi-tenant cloud data architectures.
- Estimating network and compute costs for cloud migration decisions.
- Optimizing network transfer & workflow plans across multiple clouds.
-
Abstracting the dynamic linking of heterogeneous network overlays through the use of generative programming languages.
Funder:
CHIST-ERA
Value:
€1.03m
Key outputs:
to capture device capabilities from online device documentation and automatically generate structured ontologies.
- A graphical programming tool that enables users without programming experience to easily design smart home automation workflows.
- A general ontology-based approach to describe and enable runtime reasoning and composition between different systems.
- A self-describing system that appears as a whole when viewed from above whilst potentially comprising multiple sub-systems when viewed from below.
- A platform for modelling and simulation of end-to-end IoT services and networking.
- Using LLMs to automatically compose and configure IoT services by understanding natural language descriptions of desired functionality.
- Developing cutting-edge platforms for applications and services in the areas of social connected TV, smart city services, and pervasive gaming.
Funder:
EC FP7-ICT
Value:
€20.18m
Key outputs:
- Distinctive groups of users with discernible browsing patterns that vary during the day based on analysing how 1,100+ users browse an online TV service over 6 months.
- User interaction statistics collected from a P2P multimedia service infrastructure that delivers both live and on-demand content via PC/Mac and set top box platforms.
Environmental Workbench
- Extending EVOp for larger national infrastructures, namely JASMIN.
Funder:
NERC
Value:
£1m
- A cloud-based virtual observatory for environmental sciences.
Funder:
NERC
Value:
£2m
Key outputs:
- A cloud-enabled virtual research space for different users interested in environmental science, ranging from domain specialists to the general public.
Internet Security & Resilience
- Identifying the resilience of the Internet's architecture through its fundamental building blocks.
Funder:
CPNI
Key outputs:
- Grounding key commonly known BGP issues within a structured taxonomical analysis and the first comprehensive threat model.
- Advanced network services tailored for grid applications.
Funder:
EC FP6
Value:
€2.85m
Key outputs:
- Improving grid scheduling by incorporating predictions of network performance.
- Passively measuring network performance in grid applications.