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: The Cloud Computing. Clouds allow the on-demand delivering of software, hardware, and data as services. As Cloud-based services are more numerous and dynamic, the development of efficient service provisioning policies become increasingly challenging. In this thesis we take the perspective of Software as a Service (SaaS) providers which host their applications at an Infrastructure as a Service (IaaS) provider. 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. SaaS providers want to maximize their revenues from SLAs, while minimizing the cost of use of resources supplied by the IaaS provider. Moreover, SaaS providers compete and bid for the infrastructural resources. On the other hand, the IaaS wants to maximize the revenues obtained providing virtualized resources. In this framework where multiple SaaSs have to contend the resources and where SaaSs and IaaS are in conflict, the use of Game Theory fits perfectly into the problem. The solution concept adopted is a Generalized Nash Equilibrium, when this is reached neither SaaSs nor IaaS have the interest to change strategy while the others player keeps their own strategies fixed. The analysis of this situation is not trivial, since a lot of parameters are affected by uncertainty and they are not know until run time, therefore a robust approach is adopted, i.e., the solution found remains feasible also for the worst realization of the parameters. An Analytical model is developed in order to find the equilibrium and the Gurobi solver is used to solve the game. Finally, with validation purposes, the analytical model results are compared with simulations to analyze real situations where the system stats is characterized with different and more realistic parameters distributions and limited buffer.
Negli ultimi anni il Cloud Computing è stato uno dei temi dominanti nel mondo IT, esso è semplicemente un modo per le società IT di fornire ai propri clienti software e hardware, su richiesta, tramite internet. In questo contesto interagiscono vari agenti, come i SaaS (Software as a Service) i quali forniscono software all'utente finale ospitando le loro applicazioni presso uno IaaS (Infrastructure as a Service) provider. Quest'ultimo fornisce le macchine virtuali su cui girano le applicazioni dei vari SaaS. In questa situazione conflittuale dove i SaaS vogliono minimizzare il costo delle infrastrutture e dall'altra dove lo IaaS vuole massimizzare il profitto proveniente dal noleggio delle macchine virtuali, la Teoria dei Giochi è stata applicata con successo in letteratura. In particolare, in questa tesi, il problema è stato studiato applicando il concetto di Equilibrio di Nash Generalizzato. Inoltre in questo contesto i parametri del problema sono soggetti ad incertezza, dal momento che molti sono predizioni di eventi futuri che diventano noti soltanto a run time, per questo motivo è stata studiata una formulazione robusta del problema per la quale una soluzione rimane ammissibile anche per la peggiore realizzazione dei parametri. È stato sviluppato un codice in C++ per il modello, che sfrutta Gurobi per risolvere il problema di ottimizzazione. Infine i risultati sono stati validati con l'uso di un simulatore, che simula un contesto reale e al quale sono stati applicati i risultati del modello analitico per studiarne il comportamento.
A robust game for on spot price Cloud markets
Riva, Massimiliano
2020/2021
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: The Cloud Computing. Clouds allow the on-demand delivering of software, hardware, and data as services. As Cloud-based services are more numerous and dynamic, the development of efficient service provisioning policies become increasingly challenging. In this thesis we take the perspective of Software as a Service (SaaS) providers which host their applications at an Infrastructure as a Service (IaaS) provider. 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. SaaS providers want to maximize their revenues from SLAs, while minimizing the cost of use of resources supplied by the IaaS provider. Moreover, SaaS providers compete and bid for the infrastructural resources. On the other hand, the IaaS wants to maximize the revenues obtained providing virtualized resources. In this framework where multiple SaaSs have to contend the resources and where SaaSs and IaaS are in conflict, the use of Game Theory fits perfectly into the problem. The solution concept adopted is a Generalized Nash Equilibrium, when this is reached neither SaaSs nor IaaS have the interest to change strategy while the others player keeps their own strategies fixed. The analysis of this situation is not trivial, since a lot of parameters are affected by uncertainty and they are not know until run time, therefore a robust approach is adopted, i.e., the solution found remains feasible also for the worst realization of the parameters. An Analytical model is developed in order to find the equilibrium and the Gurobi solver is used to solve the game. Finally, with validation purposes, the analytical model results are compared with simulations to analyze real situations where the system stats is characterized with different and more realistic parameters distributions and limited buffer.File | Dimensione | Formato | |
---|---|---|---|
RIVA_MASSIMILIANO_921156.pdf
non accessibile
Descrizione: Tesi
Dimensione
4.62 MB
Formato
Adobe PDF
|
4.62 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/10589/176150