Skip to main content

Algorithms and Software Systems for Energy Flexibility in Green Data Centre Using Software Defined Networking

Primary supervisor

Adel Nadjaran Toosi

Research area

Data Engineering

Interest is growing in powering data centres by energy generated from renewable sources to reduce high operational cost and carbon footprint. In 2017, Google achieved a major milestone of purchasing 100% renewable energy to match its data centres annual electricity consumption. However, efficient utilisation of renewable energy is a challenging problem due to the variable and intermittent nature of both workload demand and renewable energy supply. This is why Google underlines that 100% renewable milestone does not mean that its facilities are matched with renewable energy in every hour of every day. The goal of this proposal it to develop scheduling and resource provisioning techniques to maximize renewable energy utilisation in datacentres. To achieve the main research goal, our proposal presents innovative techniques to adapt to the workload of the data centre to match the renewable energy supply. Matching demand to supply of green energy can be performed through scheduling and resource management based on the diversity of services and applications deployed in data centres (e.g., real-time or non-real time, interactive or batch processing, best effort or guaranteed service) and their QoS requirements (e.g., delay, deadline, response time and accuracy). We aim to develop a set of novel methods such as switching off non-mandatory microservices of the application, deferred scheduling of tasks, and trimming approximation analytics in accordance with the user preference for environmental friendliness.


Required knowledge

Distributed Systems and Networking Knowledge

Software-defined networking


Learn more about minimum entry requirements.