May, 18th - May, 24th

The first version of the chapter that describes our proposed solution is finished and has been handed over to the Professor for reviewing. We are considering the possibility of submiting an article, so it is important that the chapter is reviewed and mistakes are corrected.

We have also started collecting results to present in the dissertation document. During the collection of results, we came across a use case, where a single map wave job that came after two double map wave jobs was not being launched locally, despite the fact that it had enough containers during the second wave. The job would suffer a slight delay by being launched locally, but a delay of approximately 30s (in a 10 min time window) seems like a good tradeoff when compared to bursting to the cloud and paying for the resources.

NEXT OBJECTIVES: Fix this use case and implement a tolerance to allow a certain maximum delay to avoid unnecessary bursts. Continue collecting results.