No abstract

SaaS is providing cloud applications like all the normal classes of applications of normal computing like the web applications, le applications, email applications, real-time applications, highly interactive applications, massive data analysis applications, high performance computing applications and mobile cloud applications. Cloud computing uses the internet data centers to host the applications and data storage and also the processing power in the addition to the virtualization. Clouds are a huge stack of easily and usefully virtualized services and resources (Like software, platform and hardware).These resources are used by people all over the world and dynamically congured to accommodate more and more load and scale to a huge number of users. This stack of resources and services is delivered to the customers by a pay-per-use model which veries that the services provided by the cloud providers are provided by means of service level agreements.In this master thesis, we present an assurance to the service level agreement between the cloud users and the cloud services providers by trying to assess and evaluate the cloud computing data centers services and applications provided to the cloud users. Also we introduce mathematical test models for the web, le, real time and distributed applications. We can use these test models to see the behavior of the applications over a long time.In this master thesis, we classify the cloud applications into 8 dierent classes of applications and then identify the failures and problems which aect these cloud applications.We classify also all failures types which may occur in cloud data centers like network failures, physical server failures, VM unavailability, individual failures racks and individual component failures. We got the important metrics for each class of applications and developed real scenarios for four classes. Then, we introduce dierent failures to these scenarios to see what is the impact in the most important metrics of applications. Then, we simulate these scenarios in Network Simulator 2. Finally, we solved and mitigated the failures in simulation by using two failure mitigation techniques. The rst one is the virtual machine migration and redundancy. The second technique is the forward error correction.Experimental results acquired over the course of this master's thesis give an assessment and verication of the cloud applications services provided by the cloud provider data centers. And also verify that the cloud data center provider gives their users the best services performance and high QoS, we use this verication to assure SLA response times metric and a good quality and performance of services as mentioned inside the SLA document. This verication will be given to the cloud customers.iii iv Acknowledgment "Today don't beg, don't ask, Just thank God in silence; for all the blessings in your life. " Paulo CoelhoThe road to want something is always guided by unseen strong currents of support from people who believe that you will do it what you have set...