Monday 26 September 2011

Bin Repacking Scheduling in Virtualized Datacenters: Back to work

As promised, our recent modifications in the code and in the model of the VRSP are available online. With regards to the results published in CP 2011, we improved the solving capabilities of Entropy while providing better solutions. Icing on the cake, solutions are computed about 10 times faster !

The report is available from my website or directly here.

Sunday 18 September 2011

CP 2011 is over

This week was CP 2011. Sophie was there and she presented our paper "Bin Packing Rescheduling in Virtualized Datacenters". The paper and the slides are available on my publication page.

As announced during the talk, we made some progress on the implementation since the publication and we have now better results in terms of solving duration. We plan to put them online as soon as possible.

Tuesday 6 September 2011

Mostly back to France

I am in France for a short period to sign my contract about my new position. Since September, I am officially an associate professor at OASIS! I will be back in France definitively in december, until that date, I am still working directly at Flux.

Friday 29 July 2011

PDP 2012 - Special Session on Energy-aware Systems

The submission deadline has been extended to August, 15. This special session, dedicated to energy-aware systems, will be collocated with Euromicro PDP 2012 so don't hesitate and submit !

The submission deadline for PDP 2012 has also been extended to the same date.

Friday 17 June 2011

CP 2011

I have submitted with Sophie Demassey and Xavier Lorca a paper to the application track of CP'2011. We received the acceptance notification recently.

The paper is entitled "Bin Repacking Scheduling in Virtualized Datacenters". It tries to show the interest of Constraint Programing to provide a VM manager that can be specialized on demand to consider new placement constraints. We describe a model for the problem, some placement constraints, an implementation into Entropy and some evaluations from simulated workloads.

I saw by practice, that a rebuttal process may be useful. A reviewer was not really convinced about our stuff and change its mind from our rebuttal.

The preprint version is available here.