Multicast Deployment of Cloud Operating Systems

Abstract

With cloud computing's paradigm shift for IT industries, the concept of thin clients has become popular again. There are usually numerous nodes with various functions in a cloud computing system. It is important to deploy different cloud operating systems (cloud OSs) onto different nodes. The conventional method of using unicast deployment to distribute a massive cloud OS onto thousands of nodes is time consuming and bandwidth-intensive. In this work, we propose a multicast deployment approach to significantly improve deployment efficiency. Furthermore, our multicast deployment approach can leverage existing configurations of the unicast counterpart. Therefore, customized deployment for different groups of nodes can be achieved. To evaluate the feasibility of the proposed approach in practical applications, our deployment approach is implemented using CentOS and Ubuntu on many nodes. Empirical studies show that the required time for the entire distribution process is significantly reduced, from starting delivery until the OS is ready. Moreover, network bandwidth consumption is also significantly reduced compared with the conventional unicast deployment. Consequently, system administrators must spend less effort on monitoring and maintenance.