Cloud Computing has gained a lot of popularity in the past years by offering easy access to a shared and virtualized pool of computing capabilities, emphasizing flexibility, scalability, and performance of applications. The ability to run and manage multi-clouds applications (i.e., application that runs on multiple public or private clouds) allows exploiting the peculiarities of each cloud solution and hence improves non-functional aspects such as availability, cost, and scalability. Monitoring such multi-clouds applications is fundamental to track the health of the applications and of their underlying infrastructures as well as to decide when and how to adapt their behaviour and deployment. It is clear that, not only the application but also the corresponding monitoring infrastructure should dynamically adapt in order to be optimized to the application context (e.g., adapting the frequency of monitoring to reduce network load) and to enable the co-evolution of the monitoring platform together with the multi-cloud application (e.g., if a service migrates from one provider to another, the monitoring activities have to be adapted accordingly). The goal of this work is to extend a model-based platform for the dynamic provisioning, deployment, and monitoring of multi-clouds applications to enable the dynamically adaptation of the monitoring activity that best fit with the actual deployment of the application. The model representation of the deployment simplifies the reasoning on the system and eliminates the separation between design and run time. The results achieved permit to co-evolve the monitoring activity together with the status of the running system, from the initial deployment to further modifications of the multi-cloud application. In this way, the monitoring component can autonomously monitor new resources belonging to different cloud providers and stop monitoring resources that are no longer deployed.
La popolarità del cloud computing è cresciuta esponenzialmente negli ultimi anni a causa della facilità con cui permette di accedere a risorse computazionali virtuali, enfatizzando in tal modo la flessibilità, la scalabilità e le prestazioni delle applicazioni. La possibilità di eseguire e gestire applicazioni multi-cloud (cioè applicazioni eseguite contemporaneamente su cloud diversi, pubblici o privati) permette di sfruttare le caratteristiche migliori di ogni servizio, accentuando ancora di più i vantaggi tipici del cloud computing stesso: affidabilità, riduzione costi e scalabilità. Il monitoraggio di queste applicazioni multi-cloud assume un ruolo di ri-lievo per determinare lo stato di salute delle applicazioni stesse e dell’infrastruttura sottostante. Al contempo, questa attività di monitoraggio è necessaria per decidere quando e come adattare il comportamento delle applicazioni o l’infrastruttura che le ospita. Risulta evidente come la stessa piattaforma di monitoraggio debba, a sua volta, essere adattabile per ottimizzare l’attività di monitoraggio (es. ridurre la frequenza di funzionamento per liberare risorse) e per poter coevolvere con l’applicazione (es. seguire automaticamente l’applicazione su un cloud differente). Lo scopo di questo lavoro è estendere un sistema basato su modelli per crea-re, gestire e monitorare applicazioni multi-cloud, permettendo che l’attività di monitoraggio possa essere automaticamente e dinamicamente ottimizzata in base allo stato dell’applicazione. Costruire una piattaforma basata su modelli permette inoltre di semplificare il lavoro di entità esterne che devono interagire col sistema. In aggiunta, un modello continuamente sincronizzato col sistema permette di inte-ragire su di esso anche durante l’esecuzione rompendo, di fatto, la separazione tra modellazione e mantenimento dell’infrastruttura. La piattaforma realizzata permette la coevoluzione dell’attività di monitoraggio e dell’applicazione dal momento del primo rilascio a tutte le modifiche successive, abilitando l’autoadattamento della stessa attività di monitoraggio.
Using models at runtime to support adaptable monitoring of multi-clouds applications
CIANCIARUSO, LORENZO;DI FORENZA, FRANCESCO VINCENZO
2013/2014
Abstract
Cloud Computing has gained a lot of popularity in the past years by offering easy access to a shared and virtualized pool of computing capabilities, emphasizing flexibility, scalability, and performance of applications. The ability to run and manage multi-clouds applications (i.e., application that runs on multiple public or private clouds) allows exploiting the peculiarities of each cloud solution and hence improves non-functional aspects such as availability, cost, and scalability. Monitoring such multi-clouds applications is fundamental to track the health of the applications and of their underlying infrastructures as well as to decide when and how to adapt their behaviour and deployment. It is clear that, not only the application but also the corresponding monitoring infrastructure should dynamically adapt in order to be optimized to the application context (e.g., adapting the frequency of monitoring to reduce network load) and to enable the co-evolution of the monitoring platform together with the multi-cloud application (e.g., if a service migrates from one provider to another, the monitoring activities have to be adapted accordingly). The goal of this work is to extend a model-based platform for the dynamic provisioning, deployment, and monitoring of multi-clouds applications to enable the dynamically adaptation of the monitoring activity that best fit with the actual deployment of the application. The model representation of the deployment simplifies the reasoning on the system and eliminates the separation between design and run time. The results achieved permit to co-evolve the monitoring activity together with the status of the running system, from the initial deployment to further modifications of the multi-cloud application. In this way, the monitoring component can autonomously monitor new resources belonging to different cloud providers and stop monitoring resources that are no longer deployed.File | Dimensione | Formato | |
---|---|---|---|
2014_12_Cianciaruso_diForenza.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
1.06 MB
Formato
Adobe PDF
|
1.06 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/10589/102401