In recent years the evolution and the widespread adoption of virtualization, service-oriented architectures, autonomic, and utility computing have converged letting a new paradigm to emerge: Cloud Computing. Clouds allow the on-demand delivering of software, hardware, and data as services. Currently the Cloud offer is becoming day by day wider, since all the major IT companies and Service providers, like Microsoft, Google, Amazon, HP, IBM, and VMware have started providing solutions involving this new technological paradigm. As Cloud-based services are more numerous and dynamic, the development of efficient service provisioning policies becomes increasingly challenging. In this thesis we take the perspective of Software as a Service (SaaS) providers which host their applications at multiple Infrastructure as a Service (IaaS) providers. Each SaaS needs to comply with quality of service requirements, specified in Service Level Agreement (SLA) contracts with the end-users, which determine the revenues and penalties on the basis of the achieved performance level. Each SaaS provider wants to minimize the cost of use of Cloud resources and penalties for requests execution failures. Moreover, SaaS providers compete and bid for the use of infrastructural resources. On the other hand, the IaaSs want to maximize their revenues obtained providing virtualized resources. In this thesis we model the service provisioning problem as a generalized Nash game. In particular, we develop an efficient distributed algorithm for the run-time allocation of IaaSs resources among competing SaaS providers. We demonstrate the effectiveness of our approach by performing tests considering realistic Cloud scenarios. Numerical results show that our algorithm is scalable and can be used at run-time since it can solve problem instances of maximum size in less than one minute. Compared to other state-of-the-art solutions our model can improve the efficiency of Cloud system evaluated in term of Price of Anarchy up to 100%. Furthermore our analyses point out the SaaS benefits while exploiting multiple IaaS deployment of applications and redistribution of application workloads.
Negli ultimi anni l'evoluzione e la diffusa adozione di virtualizzazione, architetture orientate ai servizi, autonomic e utility computing sono confluiti in un nuovo paradigma: il Cloud Computing. Il Cloud Computing ha come obiettivo la fornitura on-demand di software e hardware come risorse accessibili tramite Internet. Con l'aumento della quantità e della dinamicità dei servizi basati sul Cloud, lo sviluppo di politiche efficienti per la distribuzione delle risorse è diventato sempre più complesso. In questo lavoro di tesi abbiamo studiato il problema dal punto di vista dei fornitori Sofware as a Service (SaaS) che ospitano le loro applicazioni presso molteplici fornitori Infrastructure as a Service (IaaS). Ogni SaaS deve rispettare la qualità del servizio, specificata nei contratti di Service Level Agreement (SLA) con i propri clienti, che determina i ricavi e le penalità sulla base del livello di prestazioni raggiunto. Ogni SaaS vuole minimizzare i costi di utilizzo delle risorse Cloud e delle sanzioni causate dalla violazione dei contratti di SLA. Inoltre, i fornitori SaaS competono fra loro facendo offerte per l’utilizzo delle infrastrutture. Al contrario, gli IaaS vogliono massimizzare i propri introiti ottenuti fornendo risorse virtualizzate. In questa tesi abbiamo modellato il problema per la fornitura delle risorse come un gioco di Nash generalizzato. In particolare, abbiamo sviluppato un algoritmo distribuito per la gestione a run-time delle risorse degli IaaS fra i SaaS in competizione. Abbiamo dimostrato l'efficacia del nostro approccio compiendo test rappresentativi di scenari di carico reale. I risultati numerici mostrano che l'algoritmo è scalabile e può essere utilizzato a run-time, poichè può risolvere istanze del problema di dimensione massima in meno di un minuto. Rispetto ad altre soluzioni della lettaratura il nostro modello può migliorare l'efficienza in termini di Price of Anarchy del sistema Cloud valutato fino al 100%. Inoltre, le nostre analisi evidenziano i benefici che i SaaS possono ottenere sfruttando il dislocamento delle applicazioni e la distribuzione del traffico su molteplici IaaS.
Generalized Nash equilibria for the service provisioning problem in multi-Cloud systems
TREVISIOL, ETTORE
2012/2013
Abstract
In recent years the evolution and the widespread adoption of virtualization, service-oriented architectures, autonomic, and utility computing have converged letting a new paradigm to emerge: Cloud Computing. Clouds allow the on-demand delivering of software, hardware, and data as services. Currently the Cloud offer is becoming day by day wider, since all the major IT companies and Service providers, like Microsoft, Google, Amazon, HP, IBM, and VMware have started providing solutions involving this new technological paradigm. As Cloud-based services are more numerous and dynamic, the development of efficient service provisioning policies becomes increasingly challenging. In this thesis we take the perspective of Software as a Service (SaaS) providers which host their applications at multiple Infrastructure as a Service (IaaS) providers. Each SaaS needs to comply with quality of service requirements, specified in Service Level Agreement (SLA) contracts with the end-users, which determine the revenues and penalties on the basis of the achieved performance level. Each SaaS provider wants to minimize the cost of use of Cloud resources and penalties for requests execution failures. Moreover, SaaS providers compete and bid for the use of infrastructural resources. On the other hand, the IaaSs want to maximize their revenues obtained providing virtualized resources. In this thesis we model the service provisioning problem as a generalized Nash game. In particular, we develop an efficient distributed algorithm for the run-time allocation of IaaSs resources among competing SaaS providers. We demonstrate the effectiveness of our approach by performing tests considering realistic Cloud scenarios. Numerical results show that our algorithm is scalable and can be used at run-time since it can solve problem instances of maximum size in less than one minute. Compared to other state-of-the-art solutions our model can improve the efficiency of Cloud system evaluated in term of Price of Anarchy up to 100%. Furthermore our analyses point out the SaaS benefits while exploiting multiple IaaS deployment of applications and redistribution of application workloads.| File | Dimensione | Formato | |
|---|---|---|---|
|
2013_04_Trevisiol.pdf
solo utenti autorizzati dal 03/04/2016
Descrizione: Testo della tesi
Dimensione
4.35 MB
Formato
Adobe PDF
|
4.35 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/78601