Abstract
A small Cloud infrastructure for scientific computing likely operates in a saturated regime, which imposes to optimize the allocation of resources. Tenants typically pay a priori for a fraction of the overall resources. Within this business model, an advanced scheduling strategy is needed in order to optimize the data centre occupancy. FaSS, a Fair Share Scheduler service for OpenNebula, addresses this issue by satisfying resource requests according to an algorithm, which prioritizes tasks according to an initial weight and to the historical resource usage of the project. In this proceedings, we are going to describe the implementation of FaSS Version 1.0, released in March 2017 as a product of the INDIGO-DataCloud project. We are also going to discuss the results of FaSS functional and stress tests performed at the Cloud infrastructure of the INFN-Torino computing centre.
Highlights
OpenNebula (ONE) [1] is a simple and flexible open-source tool to build and manage Private Clouds
A small Cloud infrastructure for scientific computing likely operates in a saturated regime, which imposes to optimize the allocation of resources
FaSS, a Fair Share Scheduler service for OpenNebula, addresses this issue by satisfying resource requests according to an algorithm, which prioritizes tasks according to an initial weight and to the historical resource usage of the project
Summary
OpenNebula (ONE) [1] is a simple and flexible open-source tool to build and manage Private Clouds. The default OpenNebula scheduler is First-In-First-Out (FIFO) and based only on static resources partitioning among the projects This scheduling strategy applies well to a large public Cloud, where applications can scale in/out freely since the resources are approximately infinite and tenants are charged a posteriori. The FIFO scheduler, instead, is not suitable for a scientific data Cloud, for which advance resource allocation is needed, given the fact that it often operates at a saturated regime and tenants are charged a priori. For this reason, we deployed the Fair Share Scheduler (FaSS) service [2]. A demo of FaSS is available at [4]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have