Diplomarbeit DIP-2978

Shi, Qinghua: Automatic Provisioning of Sametime 3D on a Cloud Infrastructure.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Diplomarbeit Nr. 2978 (2010).
115 Seiten, englisch.
CR-Klassif.D.2.2 (Software Engineering Design Tools and Techniques)
C.2.4 (Distributed Systems)
D.2.11 (Software Engineering Software Architectures)
H.4.1 (Office Automation)
K.4.3 (Computers and Society Organizational Impacts)
KeywordsCloud; Sametime 3D; Amazon EC2; Provisioning

Virtual Collaboration for Lotus Sametime (also known as Sametime 3D) is a new collaboration tool developed by IBM Research and Development to realize real time communication, collaboration and learning in virtual worlds. A common scenario is to use it with geographically dispersed teams to conduct a brainstorming or a strategic planning session. The main benefits are seen for small teams discussing interactively and making team-based decisions. The solution uses an open-source virtual world server called OpenSimulator (OpenSim) as its platform to host virtual meetings.

A major issue in respect of providing such virtual meeting spaces on demand is that a Sametime 3D deployment cannot dynamically scale if more resources for virtual meetings are requested as currently available. Usually a physical server has to be ordered, installed and configured by an administrator. In order to solve this scalability issue an automatic provisioning system based on Cloud Computing is presented in this thesis. The automatic provisioning is realized by using provisioning flows to orchestrate different components (like connectors to Cloud providers or parts of the Sametime 3D platform) to provide virtual world server instances on demand, which can then be used by Sametime 3D. The provisioning flows are relatively simple BPEL processes which describe the steps that are necessary to provide an application. It contains steps for creating a new server instance as well as an configuration step which integrates the new server into the existing deployment topology. An important concept is the Unified Cloud Abstraction Layer which abstracts different Cloud providers, having some basic methods for managing images and instances. That enables the provisioning on different Clouds with minimal effort.

In this thesis Amazon EC2 was chosen as a Cloud provider and the integration with the provisioning flows as shown in a prototype implementation. The provisioning flows were run on an open-source BPEL engine called Apache ODE. Further it could be proved that the Cloud paradigm brings great benefit for on-demand needed computing resources: the provisioning of virtual world servers can be done efficiently and in a timely manner compared to provisioning a physical server, if an image with the virtual world server is ready on the Cloud for instantiation.

Abteilung(en)Universität Stuttgart, Institut für Architektur von Anwendungssystemen
BetreuerMatthias Hub (IBM); Mietzner, Ralph
Eingabedatum20. April 2010
   Publ. Institut   Publ. Informatik