Leveraging OpenFlow for resource placement of virtual desktop cloud applications

摘要

Popular applications such as email, photo/video galleries, and file storage are increasingly being supported by cloud platforms in residential, academia and industry communities. The next frontier for these user communities will be to transition ‘traditional desktops’ that have dedicated hardware and software configurations into ‘virtual desktop clouds’ that are accessible via thin-clients. In this paper, we describe an intelligent resource placement framework for thin-client based virtual desktops. The framework leverages principles of software-defined networking and features a ‘unified resource broker’ that uses special ‘marker packets’ for: (a) “route setup” when handling non-IP traffic between thin-client sites and data centers, (b) “path selection” and “load balancing” of virtual desktop flows to improve performance of interactive applications and video playback, and to cope with faults such as link-failures or Denial-of-Service cyber-attacks. In addition, we detail our framework implementation within a virtual desktop cloud (VDC) setup in a multi-domain Global Environment for Network Innovations (GENI) Future Internet testbed spanning backbone and access networks. We present empirical results from our experimentation that leverages OpenFlow programmable networking, as well as perfSONAR instrumentation-and-measurement capabilities for validating our framework in GENI under realistic settings. Our results demonstrate the importance of scheduling regulated measurements that can be used for intelligent resource placement decisions. Our results also show the feasibility and benefits of using OpenFlow controller applications for path selection and load balancing between thin-client sites and data centers in VDCs.