Home > Research > Publications & Outputs > Whetstone

Links

Text available via DOI:

View graph of relations

Whetstone: Reliable monitoring of cloud services

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNConference contribution/Paperpeer-review

Published
Close
Publication date18/06/2018
Host publication2018 IEEE International Conference on Smart Computing (SMARTCOMP)
PublisherIEEE
Pages115-122
Number of pages8
ISBN (electronic)9781538647059
ISBN (print)9781538647066
<mark>Original language</mark>English

Abstract

Cloud services have become powerful enablers for a variety of smart computing solutions supporting multimedia, social networking, e-commerce and critical infrastructures among others. Consequently, as we increasingly depend on the cloud, the need exists to ensure its effective role as a trustworthy services platform. Towards this objective, a plethora of cloud monitoring mechanisms have been proposed which typically assume that the collected monitoring information is reliably correct. In reality, the information collected by cloud monitors is often susceptible to reliability issues (e.g., monitor malfunctions, data corruptions, or data tampering), and obtaining reliable cloud monitoring information is still an open issue. We propose Whetstone as a novel approach to address the gap where an efficient approach of ascertaining reliable values from a set of collected monitoring data is required. To this end, Whetstone first introduces a statistical approach to filter defective data from the collected data set. Next, Whetstone develops an optimization approach to quantify the reliability of the collected data by leveraging the value deviation of the collected data. Finally, Whetstone devises a weighted aggregation approach for generating the reliable value based on the obtained information. We evaluate the proposed approach with different experimental configurations. The experimental results demonstrate the efficacy of our approach for successfully generating the maximum likelihood reliable value for raw data sets. © 2018 IEEE.