SRS Technologies VJA/HY

A Stochastic Model to Investigate Data Center Performance and QoS in IaaS Cloud Computing Systems
ABSTRACT
Cloud data center management is a key problem due to the numerous and heterogeneous strategies that can be applied, ranging from the VM placement to the federation with other clouds. Performance evaluation of Cloud Computing infrastructures is required to predict and quantify the cost-benefit of a strategy portfolio and the corresponding uality of !ervice " o!# e$perienced by users. !uch analyses are not feasible by simulation or on-the-field e$perimentation, due to the great number of parameters that have to be investigated. %n this paper, we present an analytical model, based on !tochastic &eward 'ets "!&'s#, that is both scalable to model systems composed of thousands of resources and fle$ible to represent different policies and cloud-specific strategies. !everal performance metrics are defined and evaluated to analy(e the behavior of a Cloud data center) utili(ation, availability, waiting time, and responsiveness. * resiliency analysis is also provided to take into account load bursts. +inally, a general approach is presented that, starting from the concept of system capacity, can help system managers to opportunely set the data center parameters under different working conditions.

!"ISTI#$ S%ST!M
SRS Technologies 9246669039,9246451282,9059977209,9290533483 Jagadhi.pm@gmail.com

SRS Technologies VJA/HY

D
%n order to integrate business requirements and application level needs, in terms of uality of !ervice " o!#, cloud service provisioning is regulated by !ervice ,evel *greements "!,*s#) contracts between clients and providers that e$press the price for a service, the o! levels required during the service provisioning, and the penalties associated with the !,* violations. %n such a conte$t, performance evaluation plays a key role allowing system managers to evaluate the effects of different resource management strategies on the data center functioning and to predict the corresponding costs-benefits. Cloud systems differ from traditional distributed systems. +irst of all, they are characteri(ed by a very large number of resources that can span different administrative domains. Moreover, the high level of resource abstraction allows to implement particular resource management techniques such as VM multiple$ing or VM live migration that, even if transparent to final users, have to be considered in the design of performance models in order to accurately understand the system behavior. +inally, different clouds, belonging to the same or to different organi(ations, can dynamically .oin each other to achieve a common goal, usually represented by the optimi(ation of resources utili(ation. /his mechanism, referred to as cloud federation, allows to provide and release resources on demand thus providing elastic capabilities to the whole infrastructure.

DISAD&A#TA$!S '( !"ISTI#$ S%ST!M

0n-the-field e$periments are mainly focused on the offered o!, they are

based on a black bo$ approach that makes difficult to correlate obtained data SRS Technologies 9246669039,9246451282,9059977209,9290533483 Jagadhi.pm@gmail.com

SRS Technologies VJA/HY

D
to the internal resource management strategies implemented by the system provider. !imulation does not allow to conduct comprehensive analyses of the system performance due to the great number of parameters that have to be investigated.

PR'P'S!D S%ST!M
%n this paper, we present a stochastic model, based on !tochastic &eward 'ets "!&'s#, that e$hibits the above mentioned features allowing to capture the key concepts of an %aa! cloud system. /he proposed model is scalable enough to represent systems composed of thousands of resources and it makes possible to represent both physical and virtual resources e$ploiting cloud specific concepts such as the infrastructure elasticity. 1ith respect to the e$isting literature, the innovative aspect of the present work is that a generic and comprehensive view of a cloud system is presented. ,ow level details, such as VM multiple$ing, are easily integrated with cloud based actions such as federation, allowing to investigate different mi$ed strategies. *n e$haustive set of performance metrics are defined regarding both the system provider "e.g., utili(ation# and the final users "e.g., responsiveness#.

AD&A#TA$!S '( PR'P'S!D S%ST!M

/o provide a fair comparison among different resource management strategies, also taking into account the system elasticity, a performance evaluation approach is SRS Technologies 9246669039,9246451282,9059977209,9290533483 Jagadhi.pm@gmail.com

SRS Technologies VJA/HY

D
described. !uch an approach, based on the concept of system capacity, presents a holistic view of a cloud system and it allows system managers to study the better solution with respect to an established goal and to opportunely set the system parameters.

S%ST!M ARC)IT!CT*R!

SRS Technologies 9246669039,9246451282,9059977209,9290533483 Jagadhi.pm@gmail.com

SRS Technologies VJA/HY

D

S%ST!M C'#(I$*RATI'# + )ARD,AR! C'#(I$*RATI'# +

Processor !peed &*M 9ard :isk <ey 8oard Mouse Monitor Pentium 2%V 3.3 4h( 567 M8"min# 5; 48 !tandard 1indows <eyboard /wo or /hree 8utton Mouse !V4*

SRS Technologies 9246669039,9246451282,9059977209,9290533483 Jagadhi.pm@gmail.com

SRS Technologies VJA/HY

D

S'(T,AR! C'#(I$*RATI'# +

Operating System Programming Language Java Version

: Windows XP : JAVA : JDK 1. ! a"ove.

R!(!R!#C!

:ario 8runeo, Member, %===-> * !tochastic Model to %nvestigate :ata Center Performance and o! in %aa! Cloud Computing !ystems?I!!! TRA#SACTI'#S '# PARA--!- A#D DISTRIB*T!D S%ST!MS ./012

SRS Technologies 9246669039,9246451282,9059977209,9290533483 Jagadhi.pm@gmail.com