Abstract

Hardware Virtualization is the technology that is predominantly used in cloud computing. It allows sharing of the hardware resources like CPU, RAM, Network, etc, among multiple operating system (OS) instances. Infrastructure as a service (IaaS) is a type of cloud computing which exploits the advantages of hardware virtualization. There are two categories of IaaS providers, public cloud and private cloud. Public cloud providers have a flexible cost based resources sharing system. In case of a private cloud, importance is given to the utilization of the resource available to a maximum extent, rather than limiting the resources for each OS. Hence, each virtual operating system (Guest OS) is configured with maximum possible resource and virtualizing operating system (Host OS) will dynamically allocate the resources based on request and availability. The use of dynamic resource allocation gives rise to resource crunch when one or more guest OS consume high resource due to a known or an unknown factor. In turn, this causes performance degradation of all other guest OS that are functioning normally. In this paper, we propose a software based host, guest and guest application monitoring system (VMON), that will fetch the resource demands and actual resource of Guest OS and their applications. And then generate graphs that provide the administrators a helping hand to quickly understand the bottle necks and trend in the resource utilization of each guest OS. With this information virtualization administrators can take appropriate action to prevent starvation of all other virtual machines (VM) running in that host. This is done by retrieving and storing the performance data (CPU, Memory and Disk I/O) from the hypervisor, guest OS and applications. Then the stored data is analyzed and aggregated for generating user friendly pictorial representation (multi level pie chart or rings chart). In this paper, we have included an implementation of monitoring system for VMware hypervisor ESXi 5.1. This system will plot a rings chart that will initially give a summarized view and then by clicking on the interested area, an exploded view with details is obtained.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call