Title : Scheduling optimisation of “Adaptable, Cloud, and Data Intensive” IT tasks in a renewable sources powered datacenter.
SEPIA team (IRIT laboratory) works since several years on optimisation of datacenters energy consumption.
CODIASE team (LAPLACE laboratory) work on power allocation on several electric sources and on local control of converters and backups linked to different renewable energy sources.
This PhD is funded by ANR (French National Funding Agency) in the context of DataZero project. This project is a collaboration with FEMTO (Belfort) laboratory and with EATON company (Grenoble). It focuses on optimization of datacenters using renewable energy sources.
The aim of this PhD is to optimize the planning of cloud services while taking into account electrical hazards that can occur.
The planning should provide a good resilience to services. Service scheduling will also be resilient to electrical uncertainties, adapting itself to a electrical power budget coming from the irregular renewable electricity sources.
This PhD context will be data-intensive cloud applications.
This PhD will follow the following three steps:
1) State of the art
– State of the art on renewable powered datacentrers (photovoltaic, wind, fuel cells, supercap)
– State of the art on management algorithms for renewable powered datacenters
– State of the art on energy optimization of data-intensive cloud applications
2) Expected research contributions:
To propose optimization algorithms for service cloud planing in function of a variable power budget.
To propose resilient algoritms taking into account all the characteristics of a data-intensive cloud application.
This PhD will address multi-criteria optimization problems with greedy algorithms and comparison with optimal solution on small instances.
3) Implementation and validation:
Integration in a simulation framework of:
– Data-intensive cloud applications
– Optimization and planning algorithms
– Comparison with state of the art