In this project we have developed an auto-scaling technique for an IaaS solution that takes into account the possible diverging settings of individual users. Fuzzy logic reasoning is used to process and optimise the user-specified rules. The auto-scaling engine has been tested on both Azure and AWS platforms and work is proceeding to test this with the Openstack platform, taking the Openstack ‘Heat’ component for orchestration into account. Work is ongoing on implementing a self-learning cloud controller using neural fuzzy based learning. The benefit / impact of this research is a cloud platform that can more accurately predict resource requirements, resulting in more efficient use of infrastructure and energy. A demonstration of the research results generated to date will be provided using a variable load simulator. We have one active supporter of this project but we believe that this research can be of interest to a larger number of members. A key goal of this workshop, therefore, is to explain what we have achieved to date and to understand how we can make this project more relevant for a wider subset of members. I will also provide a fuller description of the work completed in advance of the workshop.
To register to attend, please fill out form below:
