Home // CLOUD COMPUTING 2016, The Seventh International Conference on Cloud Computing, GRIDs, and Virtualization // View article

Profiling and Predicting Task Execution Time Variation of Consolidated Virtual Machines

Authors:
Maruf Ahmed
Albert Y. Zomaya

Keywords: virtualization; consolidation; performance; variation; prediction.

Abstract:
The emph{task execution time variation} (TETV) due to consolidation of emph{virtual machines} (vm), is an important issue for data centers. This work critically examines, the nature and impact of performance variation from a new angle. It introduces a simple and feasibly implementable method of using micro and syntactic benchmarks to profile vms. It is called, the emph{Incremental Consolidation Benchmarking Method} (ICBM). In this method, server resources are both individually and granularly incremented systematically, in order to quantitatively profile the effects of consolidation. Resource contention due to basic resources, like CPU, memory and I/O, have been examined separately. Extended experiments have been done on combination of those basic resources, too. All experiments have been done and data are collected from real virtualized systems, without using any simulation. The emph{least square regression} (LSR) is used on the profiled data in order to predict the TETV of vms. To profile the TETV data, the server has been consolidated with different types and levels of resource loads. The prediction process introduced here is straightforward and has low overhead, makeing it suitable to be applied on wide variety of systems. Results show that, the LSR models can reasonably well predict TETV of vms under different levels of consolidation. The emph{root mean square error} of prediction for combination of resources like, CPU-Memory, CPU-I/O and Memory-I/O are within 2.19, 2.47 and 3.08, respectively.

Pages: 103 to 112

Copyright: Copyright (c) IARIA, 2016

Publication date: March 20, 2016

Published in: conference

ISSN: 2308-4294

ISBN: 978-1-61208-460-2

Location: Rome, Italy

Dates: from March 20, 2016 to March 24, 2016