When deploying videoconferencing systems in Cloud based infrastructures, one of the most complex challenges is to distribute Multipoint Control Units (MCUs) among different servers. By addressing this challenge, we can improve the flexibility and the performance of this type of systems. However, to actually take advantage of the Cloud possibilities we have also to introduce mechanisms to dynamically schedule the distribution of the MCUs across the available resources. In this paper, we propose a resource scheduling model for videoconferencing systems and, starting from an existing MCU distribution architecture, we design a solution to enable the resource scheduling basing on custom criteria. These criteria can be based on the characteristics of each server or in their status in real time. We validate the extended model by setting up a typical videoconferencing deployment among a set of Cloud providers and testing a decision algorithm. We conclude that the proposed model enables the use of a wide range of algorithms that can be adapted to the needs of different Cloud deployments.
展开▼
