|Agrawal, Sugandha: A Service-oriented and Cloud-based Statistical Analysis Framework. |
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Masterarbeit Nr. 3670 (2014).
122 Seiten, englisch.
|CR-Klassif.||C.2.4 (Distributed Systems)|
D.2.11 (Software Engineering Software Architectures)
H.3 (Information Storage and Retrieval)
H.3.5 (Online Information Services)
Cloud Computing has gained popularity among e-Science environments after realizing the propitious use of economical provisions for delivering IT services and the range of resources offered by the cloud for the support, maintenance, and security of running the computation based applications. Cloud Computing being a recently growing technology offers various deployment and service models. In Software as a Service (SaaS) model, the applications and software run on the cloud and are available as 'pay-per-use'. As computing becomes more pervasive within the organization, the increase in complexity to manage the infrastructure of disparate architectures, distributed data and software has made computing very expensive. Cloud offerings promise to deliver all the functionality of existing information technology services at an economical cost. Researchers and scientists use resources provided by the cloud to handle large research datasets and results. The main advantages in Cloud computing are related to dynamic scaling of resources, which are able to adapt to changes based on demand of resources. Another advantage of cloud offering enables the use of multi-tenancy techniques to allow the sharing of resources between different users towards achieving the economy of scale along with considering data isolation as a dominant feature.
Representational State Transfer (REST) based architectural style has gained popularity for designing web service features like statelessness, modifiability, portability and simplicity. REST tends to focus on the components involved and their interactions along with interpretation of the significant data elements. Realising the intricacies of the computation and analysis that e-Science deals with, an attempt to provide a framework for statistical analysis has been made in this Master Thesis. The computational and numerical libraries are made available to the user and its functions provide the user with results in desirable format. Research focuses on providing such libraries can significantly and simultaneously decrease the computation time while decreasing the monetary costs of running such analyses. To enable scalability, Cloudburst technique is used to manage bursting the workload from a private cloud to public at times of capacity spikes and provide more resources on the public cloud to meet the user needs.
|PDF (2704265 Bytes)|
|Abteilung(en)||Universität Stuttgart, Institut für Architektur von Anwendungssystemen|
|Betreuer||Gómez Sáez, Santiago; Wettinger, Johannes|
|Eingabedatum||9. Dezember 2014|