An adaptive latency mitigation scheme for massively multiuser virtual environments

  • Authors:

  • Behnoosh Hariri;Shervin Shirmohammadi;Mohammad Reza Pakravan;Mohammad Hossein Alavi

  • Affiliations:

  • Distributed Collaborative Virtual Environment Research (DISCOVER) Lab, School of Information Technology and Engineering, University of Ottawa, 800 King Edward Ave., Ottawa, Ontario, Canada K1N 6N5 ...;Distributed Collaborative Virtual Environment Research (DISCOVER) Lab, School of Information Technology and Engineering, University of Ottawa, 800 King Edward Ave., Ottawa, Ontario, Canada K1N 6N5;Advanced Communications Research Institute (ACRI), Department of Electrical Engineering, Sharif University of Technology, PO Box 11365-8639, Azadi Ave., Tehran, Iran;Advanced Communications Research Institute (ACRI), Department of Electrical Engineering, Sharif University of Technology, PO Box 11365-8639, Azadi Ave., Tehran, Iran

  • Venue:
  • Journal of Network and Computer Applications
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

As massively multiuser virtual environments (MMVEs) expand in terms of size and user population, they tend toward using P2P architectures as a way to provide scalability without the need for large centralized resources. Distributed hash table (DHT)-based networks have been introduced as a promising option for overlay-based distributed massively multiuser virtual environment applications. However, overlay latency stretch seriously affects MMVE performance where QoS is crucial for real-time user collaboration. This work includes a series of efforts in the alleviation of such undesired latency. Our approach to latency mitigation consists of two phases. First, we propose a position-based ID assignment approach to minimize message hop-count by exploiting the clustered pattern of traffic exchange among MMVE users. Second, we introduce a new ant-based distributed neighbor selection scheme that can be used by MMVE users to select the best neighbors within their areas of interest. In order to evaluate the performance of this heuristic approach, we model the neighbor selection problem in the form of a network flow problem and use its solution as an optimality bound to compare the results. Simulation results demonstrate that the proposed algorithms will compensate for DHT latency stretch to a high extent and the performance of the resulting system would closely follow the optimal bound while communication overhead is negligible.