Home // CLOUD COMPUTING 2017, The Eighth International Conference on Cloud Computing, GRIDs, and Virtualization // View article

On Exploiting Resource Diversity in the Public Cloud for Modeling Application Performance

Authors:
Mark Meredith
Bhuvan Urgaonkar

Keywords: public cloud; tenant workload; performance modeling

Abstract:
Cloud computing platforms, such as Amazon EC2, Google Computing Engine, and Microsoft Azure, offer dozens of virtual machine (VM) types with a wide range of resource capacity vs. price trade-offs, requiring a customer to consider numerous resource configurations when evaluating service needs. We investigate the possibility of exploiting this diversity of VM types to predict the performance of workloads on new VM types using black box modeling. The performance model used is a multiple linear regression of the average application response time as a function of VM load (throughput in requests per second), the number of CPU cores, and main memory capacity. For three different types of data storage applications - Redis (key-value stores), Apache Cassandra (a NoSQL database) and MySQL (an ACID database) - the model accuracy improves when the training data spans more diverse VMs. E.g., for Redis, the R2_predicted measure of model efficacy improves from 0.4-0.5 with 2 VM types for training and 0.7 for 3 VM types to 0.8 for 4 VM types. These results suggest further interesting research challenges, such as the possibility of automating the process of calibrating performance models using diverse resource types on a public cloud leading to ``performance modeling as a service.''

Pages: 66 to 72

Copyright: Copyright (c) IARIA, 2017

Publication date: February 19, 2017

Published in: conference

ISSN: 2308-4294

ISBN: 978-1-61208-529-6

Location: Athens, Greece

Dates: from February 19, 2017 to February 23, 2017