A Cross-Layer Approach-Based Gnutella for Collaborative Virtual Environments over Mobile Ad Hoc Networks
A Cross-Layer Approach-Based Gnutella for Collaborative Virtual Environments over Mobile Ad Hoc Networks
复制标题
DOI：
10.1109/tpds.2009.91
复制
发表时间：
2010-07
影响因子：
5.3
通讯作者：
A. Boukerche;Anis Zarrad;R. B. Araujo
中科院分区：
文献类型：
--
作者：
A. Boukerche;Anis Zarrad;R. B. Araujo
协同虚拟环境（Collaborative Virtual Environments，CVE），如军事训练、应急准备模拟和在线游戏，当参与的用户通过他们的移动的设备在ad hoc网络中共享3D虚拟环境时，对网络性能提出了严格的要求。然而，现有CVE的大多数支持系统都局限于桌面设置，从而阻止了用户在移动的时的协作。在这项工作中，我们提出并评估了一个Gnutella对等网络在移动的ad hoc网络（MANDIX），以支持移动的协作虚拟环境（MCVE）的应用。这两种架构都具有一些共同的特征，例如去中心化、自重组和自修复。然而，对等网络和移动自组网在不同的网络层上运行，因此导致性能较差。为了解决这个问题，我们探索了一种新的跨层方法，以提高整体性能的CVE在移动自组网。网络层应该知道用户在虚拟环境（VE）中的物理位置，以最大限度地减少网络流量，并科普节点之间的适度工作负载。本文提出了一种基于移动自组网的协同虚拟环境的跨层方法模型。我们描述了它的实现，并使用NS2模拟器评估其性能。
Collaborative Virtual Environments (CVEs) such as military training programs, emergency preparedness simulations, and online games place strict requirements on network performance when participating users share the 3D virtual environment through their mobile devices in an ad hoc network. However, most support systems for existing CVEs are confined to a desktop setting, thus preventing collaboration while users are mobile. In this work, we propose and evaluate a Gnutella peer-to-peer network over mobile ad hoc networks (MANETs) to support Mobile Collaborative Virtual Environment (MCVE) applications. Both architectures share some common features such as decentralization, self-reorganization, and self-healing. However, peer-to-peer networks and MANETs operate on different network layers, and thus, cause poor performance. To address this problem, we explore a novel cross-layer approach to improve the overall performance of CVEs over MANETs. The network layer should be aware of the user's physical position in the virtual environment (VE) in order to minimize network traffic and cope with a moderate workload between nodes. In this paper, we present a cross-layer approach model for collaborative virtual environment over MANET. We describe its implementation and evaluate its performance using NS2 simulator.