Application-Aware Infrastructure Clustering for Cloud Service Placement
to Enhance User QoE
D. Chemodanov, and P. Calyam. QCMan 2016 (QCMan 2016), Würzburg, Germany, (September 2016)
Abstract
Cloud services placement can be suboptimal in certain cases due to
inefficient network design, which in turn impacts user Quality of
Experience (QoE). Consequently, Application Service Providers (ASPs) need
to manage cloud network infrastructures with efficient designs that fully
satisfy Service Level Objectives (SLOs) of their data/video-intensive
applications of users. To proactively avoid this problem, a straightforward
solution used by ASPs is to have many replicas of their services by renting
more resources from Infrastructure Providers (InPs) which can lead to an
expensive service delivery proposition. In this paper, we present a novel
possibilistic C-Means (PCM) approach to enhance user QoE in cloud service
placement by clustering network infrastructure with awareness of user SLO
satisfaction amidst network path constraints. Our evaluation results
obtained using numerical simulations as well as in a real-world cloud
testbed with actual users prove that our multi-constrained path aware PCM
approach outperforms existing solutions. Specifically, we show how our
proposed infrastructure clustering with the PCM approach allows ASPs to
rent less resources from InPs that reduces user cost, while still
delivering satisfactory user QoE.
%0 Conference Paper
%1 Chemodanov2016
%A Chemodanov, Dmitrii
%A Calyam, Prasad
%B QCMan 2016 (QCMan 2016)
%C Würzburg, Germany
%D 2016
%K itc itc28
%T Application-Aware Infrastructure Clustering for Cloud Service Placement
to Enhance User QoE
%U https://gitlab2.informatik.uni-wuerzburg.de/itc-conference/itc-conference-public/-/raw/master/itc28/Chemodanov2016.pdf?inline=true
%X Cloud services placement can be suboptimal in certain cases due to
inefficient network design, which in turn impacts user Quality of
Experience (QoE). Consequently, Application Service Providers (ASPs) need
to manage cloud network infrastructures with efficient designs that fully
satisfy Service Level Objectives (SLOs) of their data/video-intensive
applications of users. To proactively avoid this problem, a straightforward
solution used by ASPs is to have many replicas of their services by renting
more resources from Infrastructure Providers (InPs) which can lead to an
expensive service delivery proposition. In this paper, we present a novel
possibilistic C-Means (PCM) approach to enhance user QoE in cloud service
placement by clustering network infrastructure with awareness of user SLO
satisfaction amidst network path constraints. Our evaluation results
obtained using numerical simulations as well as in a real-world cloud
testbed with actual users prove that our multi-constrained path aware PCM
approach outperforms existing solutions. Specifically, we show how our
proposed infrastructure clustering with the PCM approach allows ASPs to
rent less resources from InPs that reduces user cost, while still
delivering satisfactory user QoE.
@inproceedings{Chemodanov2016,
abstract = {Cloud services placement can be suboptimal in certain cases due to
inefficient network design, which in turn impacts user Quality of
Experience (QoE). Consequently, Application Service Providers (ASPs) need
to manage cloud network infrastructures with efficient designs that fully
satisfy Service Level Objectives (SLOs) of their data/video-intensive
applications of users. To proactively avoid this problem, a straightforward
solution used by ASPs is to have many replicas of their services by renting
more resources from Infrastructure Providers (InPs) which can lead to an
expensive service delivery proposition. In this paper, we present a novel
possibilistic C-Means (PCM) approach to enhance user QoE in cloud service
placement by clustering network infrastructure with awareness of user SLO
satisfaction amidst network path constraints. Our evaluation results
obtained using numerical simulations as well as in a real-world cloud
testbed with actual users prove that our multi-constrained path aware PCM
approach outperforms existing solutions. Specifically, we show how our
proposed infrastructure clustering with the PCM approach allows ASPs to
rent less resources from InPs that reduces user cost, while still
delivering satisfactory user QoE.},
added-at = {2016-08-31T16:30:53.000+0200},
address = {Würzburg, Germany},
author = {Chemodanov, Dmitrii and Calyam, Prasad},
biburl = {https://www.bibsonomy.org/bibtex/20c653594a51cf92a6bde14ea8f0706c4/itc},
booktitle = {QCMan 2016 (QCMan 2016)},
days = {12},
interhash = {fa9be042107af9e4a4a0cd116fb6b95f},
intrahash = {0c653594a51cf92a6bde14ea8f0706c4},
keywords = {itc itc28},
month = {Sept},
timestamp = {2020-05-26T16:53:35.000+0200},
title = {Application-Aware Infrastructure Clustering for Cloud Service Placement
to Enhance User QoE},
url = {https://gitlab2.informatik.uni-wuerzburg.de/itc-conference/itc-conference-public/-/raw/master/itc28/Chemodanov2016.pdf?inline=true},
year = 2016
}