Evaluation of Scheduling Techniques in Cloud Data Centers: Simulation and Experimentation

Abstract

Cloud Computing technology is becoming a challenging technology requirement for many enterprises and individuals nowadays. Due to its numerous advantages offered to its clients such as elasticity, scalability and cost-effectiveness the cloud computing industry has grown much since it emerged and is foreseen to continue this growth through the coming years. This growth requires cloud providers to expand the size and number of their data centers to accommodate the needs of their users. As the number of data centers increases, efficient scheduling algorithms are required in order to manage the scheduling of arrivals to a data center where it can be served optimally. This paper discusses two proposed scheduling algorithms for two distributed cloud data centers, one for cases of real-time requests with bounded delays and another for data centers having diverse geographical locations. An OMNeT++ simulation model is introduced as well as a test-bed implementation in order to test the proposed algorithms and their effectiveness under general assumptions and more realistic conditions.