Summary
Scheduling plays an important role in distributed computing. However, there is a well recognised gap between Scheduling Theory and Distributed Computing Practice. Below are two examples of research topics aimed at bridging that gap.
Optimisation algorithms for task offloading in Edge computing should perform decision making for a combined, Edge-Cloud system. In a typical scenario, users submit tasks to Edge servers. Some tasks can be processed by the Edge, which capacity is limited, and the results are then returned back to users. Remaining tasks are moved to the Cloud, with abundant capacity and typically with faster servers, but there are additional costs and time incurred by data transfer. Thus, a number of a challenging optimsation problems arise aimed at finding solutions which minimise task completion times, computation and data transfer costs, and maximise the system throughput rate.
Task consolidation algorithms in cloud computing systems is another research area, which goal is to improve energy efficiency of cloud computing systems. The underlying optimisation problem, known as bin packing, is notoriously hard to solve, especially in its enhanced form, which reflects features of various distributed computing scenarios. The project will include modelling of those features, developing new algorithms which enrich existing libraries of bin packing algorithms, algorithm implementation and evaluation.
Familiarity with optimisation algorithms and good programming skills are preferred.
