Masterarbeit MSTR-2016-33

Mohamed, Sherif Yassin Mostafa: Evaluation of Alternative Distributions of a Choreography-based Smart City Transportation Application among Cloud Offerings.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Masterarbeit Nr. 33 (2016).
117 Seiten, englisch.
CR-Klassif.C.4 (Performance of Systems)
D.2.5 (Software Engineering Testing and Debugging)
D.2.8 (Software Engineering Metrics)
D.4.8 (Operating Systems Performance)

In the last years, the amount of applications that are partially or completely migrated to cloud environments has significantly increased. Cloud providers offer nowadays multitude of services, following the Everything-as-a-Service (*aaS) model. However, the utilization of the cloud services involved in such migrations can have high impact over the application’s performance and cost effectiveness. Choreography-based applications, i.e., applications modeled as service choreographies, are typical candidates, due to its distributed nature. In the scope of this thesis we use the Urban Mobility application, which is a choreographybased application introduced in the EU project ALLOW Ensembles. The application lies within the domain of Smart Cities, targeting the utilization and adaptation of computation & technologies in order to offer citizens mobility and transportation services in an efficient manner. The work done in this thesis relies on the decision to migrate the Urban Mobility application to the cloud, i.e., to cloudify its application stack. Towards such a goal, in the thesis we evaluate a set of cloud services spanned among multiple cloud providers. In particular, we analyze and derive a set of viable distribution alternatives of the application components, using different cloud services, such as IaaS or DBaaS. Finally this thesis designs and drives an extensive empirical analysis to evaluate the impact of different distribution alternatives, focusing on the performance and cost offered by the different cloud services. The problem tackled is mutli-dimensional. There is no "optimal" distribution with respect to all evaluated aspects. Nevertheless, the empirical and analytical results produced in this thesis are aimed to report the performance of different application viable distributions, which can be leveraged in the decision making tasks related to the selection and configuration of cloud services.

Volltext und
andere Links
PDF (6689966 Bytes)
Zugriff auf studentische Arbeiten aufgrund vorherrschender Datenschutzbestimmungen nur innerhalb der Fakultät möglich
Abteilung(en)Universität Stuttgart, Institut für Architektur von Anwendungssystemen
BetreuerLeymann, Prof. Frank; Gómez Sáez, Santiago
Eingabedatum1. August 2018
   Publ. Institut   Publ. Informatik