Åbo universitet

Capstone

The major ICT companies, such as Google, Microsoft, Amazon, and Facebook are operating large data centers around the world for cloud computing and application hosting. The growing user demand of cloud computing has considerably increased the energy consumption of data centers, which has become a critical issue. High energy consumption not only translates to the high cost, but also leads to high carbon emissions which are not environmentally friendly. There is a growing interest in “Green Data Centers” and there is still a big gap in this area to be covered. Dynamic VM consolidation through live migration is a popular way towards energy conservation by packing Virtual Machines (VMs) on to the minimum number of active servers. However, achieving the desired level of Quality of Service (QoS) between user and a data center is critical. Therefore, the design of VM consolidation methods is challenging (NP-hard problem) and dynamic.

This project presented two heuristic algorithms for dynamic VM consolidation. These algorithms reduce the energy consumption of data centers by consolidating VMs into a reduced number of active Physical Machines (PMs) while preserving Quality of Service (QoS) requirements. To evaluate the efficiency of our proposed methods, we set up experimental environment using the CloudSim toolkit. CloudSim is a powerful discrete event simulator for implementation and evaluation of VM consolidation techniques for different applications. We evaluated the performance of our proposed approaches by conducting experiments with ten different real workload traces. Furthermore, we use OpenStack Cloud platform to evaluate the proposed VM consolidation algorithms in a real cloud environment. Generally, this project tackles the research challenges in relation to energy-efficient dynamic VM consolidation in data center under QoS constraints. In particular, following research problems are investigated:

  • When is the best time to migrate VMs from an overloaded or under-loaded server?
  • Which VMs to migrate that can reduce the energy consumption and improve QoS ?
  • Where to migrate the VMs selected for migration?
  • When and which server to switch on/off?