Master Thesis MSTR-3567

BibliographyDas, Darsana: Integrating Cloud Service Deployment Automation With Software Defined Environments.
University of Stuttgart, Faculty of Computer Science, Electrical Engineering, and Information Technology, Master Thesis No. 3567 (2014).
62 pages, english.
CR-SchemaD.2.13 (Software Engineering Reusable Software)
K.6 (Management of Computing and Information Systems)
Abstract

The last decade has seen a deluge of technology and services related to cloud -computing. The idea of traditional IT infrastructure has been usurped by a rush to move all services to a cloud infrastructure and to implement them in a cloud-native way.

A cloud infrastructure is unlike a conventional infrastructure in that, it is implemented as a Software-Defined Environment (SDE) which abstracts and virtualises the underlying physical resources. It manages physical resources and provides virtual resources that process, manage, store, create networks and provision services. But while the concept of a virtual data center as played out by an SDE seems ideal, the issues it brings in implementation are quite a few. For one, deployment and management of services on such a large scale is difficult. It is also subject to human efficiency and error. So, new levels of tracking and automation have to be explored in parallel to support and sustain the cloud-computing phenomenon.

Several configuration management and orchestration tools have been developed over years, to fulfill the need of automated deployment and management. TOSCA is one of such that provides a specification which empowers this goal by using a metamodel. It defines implementation artifacts that encapsulate functionality of external services to use them during deployment. An open-source container called OpenTOSCA, is under development at the University of Stuttgart. This framework can process the TOSCA specification.

In this thesis, OpenStack has been employed as the cloud infrastructure provider. To achieve our goal, an implementation artifact for the relevant OpenStack components has to be developed in concept and implementation. In order to create useful implementation artifacts an appropriate level of abstraction has to be introduced between the low-level OpenStack API and the services that the implementation artifact provides. Accordingly, a TOSCA artifact is designed and implemented as a web service and packaged in a container file, called CSAR. This file, when processed by the OpenTOSCA container, causes the automated provisioning of a server on OpenStack. To evaluate the nature of the implementation artifact, the web service is hosted on the local machine and tests are performed to check for correct implementation of service endpoints. The results of these invocations are later presented.

Full text and
other links
PDF (2738015 Bytes)
Department(s)University of Stuttgart, Institute of Parallel and Distributed Systems, Parallel Systems
Superviser(s)Waizenegger, Tim
Entry dateMay 5, 2014
   Publ. Department   Publ. Institute   Publ. Computer Science