Analysis of Virtualization Technologies for High Performance Computing Environments

Abstract

Abstract-As Cloud computing emerges as a dominant paradigm in distributed systems, it is important to fully understand the underlying technologies that make clouds possible. One technology, and perhaps the most important, is virtualization. Recently virtualization, through the use of hypervisors, has become widely used and well understood by many. However, there are a large spread of different hypervisors, each with their own advantages and disadvantages. This manuscript provides an in-depth analysis of some of today's commonly accepted virtualization technologies from feature comparison to performance analysis, focusing on the applicability to High Performance Computing environments using FutureGrid resources. The results indicate virtualization sometimes introduces slight performance impacts depending on the hypervisor type, however the benefits of such technologies are profound and not all virtualization technologies are equal.

A Survey on Different Method of Balancing a Binary Search Tree

“…Although it is known that if input is random, we will be closer to a balanced tree [17] [18]. Still some balancing technique is required to prevent the tree from becoming higher on one side resulting after a series of insertions and deletions [19] [20].…”

A Holistic Model of the Performance and the Energy-Efficiency of Hypervisors in an HPC Environment

“…While the second point is now circumvented on the considered hypervisor systems, the first one remains challenging. Another study that used to guide not only our benchmarking strategy but also our experimental setup is the evaluation mentioned in the previous section that was performed on the FutureGrid platform [7]. The targeted hypervisors were Xen, KVM, and Virtual Box and a serious performance analysis is proposed, with the conclusion that KVM is the best overall choice for use within HPC Cloud environment.…”

“…For reproducibility reasons, all of them are open source and we based our choice on a previous study operated in the context of the FutureGrid † platform [7], and a better focus on I/O operation that we consider as under-estimated in too many studies involving virtualization evaluation. We thus arrived to the three benchmarks:…”

Virtual Machines Detection Methods Using IP Timestamps Pattern Characteristic

“…For comparison reasons, VMs and a real machine were used as the target machine. The VMs were popular virtualization products for open platforms [9], i.e., VMWare [10], Oracle VirtualBox [11] and Xen [12], were used as the VM target machine's hypervisors, and the real machine ran open source Ubuntu Linux as the native operating system. Our findings show that by analyzing the pattern of the IP timestamps in the replies from the target machines, we were able to distinguish between the VM and real machine environments.…”