Home // CLOUD COMPUTING 2019, The Tenth International Conference on Cloud Computing, GRIDs, and Virtualization // View article

Efficient Virtual Machine Consolidation Approach Based on User Inactivity Detection

Authors:
Jan Fesl
Vineet Gokhale
Marie Feslova

Keywords: cloud; consolidation; neural network

Abstract:
Large cloud architectures consist of numerous high-performance servers, each hosting a multitude of Virtual Machines (VMs). Naturally, the server resources for processing and storage are shared among VMs, which, in turn, could be simultaneously accessed by several authorized users. Resource reallocation takes place after a session terminates. However, failure of systematic session termination causes blockage of resources resulting in severe under-utilization. In order to mitigate such scenarios, one needs to efficiently detect user inactivity for timely release of the resources. This is a non-trivial task. To this end, we propose a hybrid resource-desktop monitoring technique, which involves capturing of user interaction with the client computer, in addition to monitoring the client-server network activity. The rationale behind this approach is that even in case of lightweight applications, the user interactions cause continuous changes in the visual contents being displayed. Periodic screenshots of the client screen and network activity between client and server provide crucial information about the user inactivity. Our preliminary investigation suggests that such self-organizing virtualization infrastructure is a promising direction for the design of modern cloud-based services.

Pages: 105 to 109

Copyright: Copyright (c) IARIA, 2019

Publication date: May 5, 2019

Published in: conference

ISSN: 2308-4294

ISBN: 978-1-61208-703-0

Location: Venice, Italy

Dates: from May 5, 2019 to May 9, 2019