A generic framework for application streaming service

The Cloud provides a shared computing environment to execute complex and compute intensive virtual appliances (VA). With network function virtualization and software defined networks, we are expected to see the physical network devices and their management consoles as VAs streamed on demand. Here, our focus is on the download mechanism that is used to distribute software that requires complex installation and rigid dependencies on the operating system. With limited bandwidth, it is difficult to download several pieces of software, especially with slow or sporadic internet connectivity. In this paper, we present a novel framework for application streaming named as Ceaseless Virtual Appliance Streaming (CVAS) system that offers virtual machine streaming through small executable blocks. The user can use a VA while it is being downloaded. The experimental section demonstrates the effectiveness of the proposed framework in terms of buffering time to execute virtual appliances. A feature analysis in comparison to competing approaches demonstrates that CVAS has improved buffering time, responsiveness and state pre-fetching accuracy.