Skip lists: a probabilistic alternative to balanced trees
Communications of the ACM
Looking up data in P2P systems
Communications of the ACM
Chord: a scalable peer-to-peer lookup protocol for internet applications
IEEE/ACM Transactions on Networking (TON)
Evaluation of a Pre-Reckoning Algorithm for Distributed Virtual Environments
ICPADS '04 Proceedings of the Parallel and Distributed Systems, Tenth International Conference
An Architecture for Web-Services Based Interest Management in Real Time Distributed Simulation
DS-RT '04 Proceedings of the 8th IEEE International Symposium on Distributed Simulation and Real-Time Applications
Proceedings of the 20th Workshop on Principles of Advanced and Distributed Simulation
Hi-index | 0.00 |
Interest Management (IM) schemes for Distributed Virtual Environments (DVEs) provide a means to scale-up the number of participants and objects in a DVE instance by reducing the amount of information DVE components send and/or receive across a communication network and reducing the amount of DVE state information a DVE component must keep. In this paper we propose a model for propagating formula interest expressions that does not use intermediary servers. This work allows DVEs to be constructed where the content that composes a virtual world is provided by DVE components and DVE components are free to join and leave a DVE instance. Content may be added and/or removed dynamically (during run-time) from a DVE instance. Communication is carried out via a partially interconnected unicast peer-to-peer network using sender-based filtering. The proposed model determines the DVE components capable of evaluating a formula Interest Expression (IE). A formula IE is only sent to these DVE components. The proposed model employs a two-tiered distributed index of Interest Operator (IO) information to determine the DVE components capable of evaluating a formula IE. The first tier is used to identify the DVE components capable of evaluating a formula IE. The second tier indexes the first tier. This paper describes the logical operation and functional building blocks of the model and presents a high level analysis of its performance in comparison with an existing simple model.