Scalable propagation of continuous actions in peer-to-peer-based massively multiuser virtual environments : the continuous events approach

Heger, Florian

Vorschau

PDF
diss_fheger.pdf - Veröffentlichte Version
Download (5MB)

URL:	https://madoc.bib.uni-mannheim.de/33234
URN:	urn:nbn:de:bsz:180-madoc-332341
Dokumenttyp:	Dissertation
Erscheinungsjahr:	2013
Ort der Veröffentlichung:	Mannheim
Hochschule:	Universität Mannheim
Gutachter:	Becker, Christian
Datum der mündl. Prüfung:	29 April 2013
Sprache der Veröffentlichung:	Englisch
Einrichtung:	Fakultät für Betriebswirtschaftslehre > Wirtschaftsinformatik II (Becker 2006-2021)
Fachgebiet:	004 Informatik
Normierte Schlagwörter (SWD):	Peer-to-Peer-Netz , Virtuelle Realität , Skalierbarkeit
Freie Schlagwörter (Englisch):	Peer-to-Peer , MMVEs , Scalability
Abstract:	Peer-to-Peer-based Massively Multiuser Virtual Environments (P2P-MMVEs) provide a shared virtual environment for up to several thousand simultaneous users based on a peer-to-peer network. Users interact in the virtual environment by controlling virtual representations of themselves, so-called avatars. Their computers communicate with each other via a wide area network such as the Internet to provide the shared virtual environment. A crucial challenge for P2P-MMVEs is propagating state changes of objects in the virtual environment between a large number of user computers in a scalable way. Objects may change their state on one of the computers, e.g. their position. Information about a state change has to be propagated via the peer-to-peer network to computers of other users whose avatars are able to perceive the object. Optimization algorithms for a scalable propagation of state changes are needed because of the very large number of users and the typically limited bandwidth of their Internet connections. This thesis describes an approach that optimizes the propagation of state changes caused by continuous actions. Continuous actions lead to multiple subsequent state changes over a given period of time. Instead of propagating each subsequent state change caused by continuous actions via the network, the approach propagates descriptions of the actions included in so-called continuous events. Based on the descriptions, the subsequent state changes are calculated and applied over time on each user's computer. Continuous events contain information about (1) the timing of calculations, (2) the spatial extent of the influence of the continuous action in the virtual environment over time and (3) the effect of the continuous action on influenced objects over time. The propagation and management of continuous events is performed based on the spatial publish subscribe communication model. Each user computer declares interest in a certain space in the virtual environment. If the space intersects with the spatial extent of the influence of a continuous event, the particular computer is provided with the continuous event. This thesis describes the basic concept of continuous events, presents a system architecture for support of continuous events in the context of a given target system model for P2P-MMVEs, and evaluates the continuous events approach based on a prototypical implementation of the system architecture.