Copyright (c) 2009 All rights reserved. http://works.bepress.com/pboustead Recent documents in Dr. Paul Boustead en-us Sun, 31 May 2009 10:21:27 PDT 3600 http://works.bepress.com/pboustead/26 http://works.bepress.com/pboustead/26 Thu, 29 Jan 2009 21:10:26 PST Scribe is a scalable application level multicast infrastructure. We have developed two techniques to improve the performance of Scribe in terms of latency and bandwidth distribution. The first technique identifies that the final hop of Scribe traffic path is largely selected without any proximity consideration and incurs the longest distance traveled. To overcome this, we introduce Proximity Neighbor Selection (PNS) into the final hop for latency improvement. The second technique builds a hierarchical two-level overlay. While PNS can be applied at both levels for latency performance, the bandwidth stress required by applications can now be distributed among the nodes in the higher level overlay. Our simulation using GT-ITM topology has shown that both techniques have improved the latency performance for more than 30 percent, and the two-level overlay has improved the bandwidth distribution up to 2.7 times, comparing with what can be achieved by a standard Scribe overlay. We have developed the techniques in the context of Massively Multiplayer Online Games (MMOGs). While Scribe provides a possible platform for the scalable deployment of MMOGs, game developers may leverage the techniques to enhance the design of real-time interactions between players in the game world. X. Jiang http://works.bepress.com/pboustead/25 http://works.bepress.com/pboustead/25 Mon, 02 Jun 2008 23:06:05 PDT As a future communication tool to the Distributed Virtual Environment, the Immersive VoIP service delivers to each user a real-time mix of nearby user voices which are all rendered according to their respective virtual world positions. In this work, we conjecture that the voice latency constraints between Immersive VolP users would vary according to their virtual world distance of separation. As verified in our subjective listening tests, this conjecture is more realistic than fixed latency constraint, in terms of user perceptual experience and network feasibility. Empowered by the Bandwidth Reduced server assignment algorithm, the proposed two-hop distributed server architecture achieves a better balance between access bandwidth cost and voice latency when compared against the central server and the peer-to-peer architectures. The two-hop architecture targets the challenging scenario where there lacks clear correlations between the virtual world and the underlying physical network. In addition, we have also investigated the necessary frequency to execute the server assignment solutions to mitigate sensitivity to virtual world mobility. Y. P. Que http://works.bepress.com/pboustead/24 http://works.bepress.com/pboustead/24 Thu, 29 May 2008 09:48:25 PDT Information dissemination (flooding) forms an integral part of routing protocols, network management, service discovery and information collection (sensing). Given the broadcast nature of ad hoc network communications, information dissemination provides a challenging problem. This paper introduces Utility Based Flooding (UBF). UBF is a distributed optimised flooding mechanism for ad hoc networks that unlike existing optimised flooding algorithms is fully resource aware. Resource awareness is achieved by assigning a forwarding utility to neighouring nodes to determining the desirability of a neighbouring node in continuing a flood. UBF is particularly applicable to ad hoc network environments composed of heterogeneous nodes that may have varying characteristics and constraints. In this paper, UBF is compared to existing flooding mechanism in a constrained environment. Nodes are assigned varying degrees of remaining battery power and user based constraints that limit a nodes benevolence based upon its remaining battery power. We show through simulation that UBF compared to Utility Based Multipoint Relay (UMPR) flooding, Multipoint Relay (MPR) flooding and Blind flooding significantly improves broadcast reachability over successive broadcasts, does not adversely affect performance and extends the lifetime of the network. UBF delivers packets to over 90% of nodes in the network for over 70 successive broadcasts. Blind flooding, UMPR and MPR are only able achieve 42, 39 and 23 successive broadcasts respectively. J. Lipman http://works.bepress.com/pboustead/23 http://works.bepress.com/pboustead/23 Thu, 29 May 2008 09:48:22 PDT The popularity of on-line games that emphasize real-time interactivity is on the rise. A survey of issues and techniques for supporting on-line games is provided in this paper. Latency and scalability are two primary aspects on which this survey is based. Network latency heavily influences the design of on-line games with regards to the nature of interactivities involved, as well as, real-time flows such as voice and video if enhanced user immersive experiences are required. As a consequence, a trade-off between consistency and responsiveness, the two latency-related artifacts, is usually needed. This can be demonstrated through a number of latency compensation techniques presented in this paper, including the synchronization and optimistic approaches. To address the drawbacks of the current predominant client-server model, scalable architectures usually achieve scalability through the partition of either the physical or virtual world. In the case of distributed server design, this has resulted in systems such as the locale or proxy server architecture. In the peer-to-peer design, in addition to partition, hybrid system introduces the concept of 'super node' to allow the manageability in the overlay network. The latest peer-to-peer design emphasizes a structured approach such as the Pastry. Although such a design is not targeted for latency constraint applications, due to its unique features, experiments on hosting networked games on structured peer-to-peer network start to emerge. X. Jiang http://works.bepress.com/pboustead/22 http://works.bepress.com/pboustead/22 Thu, 29 May 2008 09:48:19 PDT Optical burst switches (OBSes) have been proposed to improve the utilization of a network of optical cross connect (OXCs). Current studies on OBS assume a network consisting of OBSes alone. While this is a reasonable assumption for evaluating a new technology, the question of how a network of OXCs can be evolved to a network of OBSes has not been studied. In this paper, we propose a hybrid architecture consisting of OBSes at the network edge and OXCs in the network core. This architecture allows carriers to gradually migrate from an OXC-based network to an OBS-based network with an improved network utilization. In addition, we use queueing analysis to study the performance of this new architecture. C. T. Chou http://works.bepress.com/pboustead/21 http://works.bepress.com/pboustead/21 Thu, 29 May 2008 09:48:16 PDT The available bandwidth in an ad-hoc network can decrease rapidly due to the shared access to the communication channel. Power control can enhance the available bandwidth by allowing non-interfering nodes to communicate simultaneously. A number of power aware topology control algorithms have been proposed, where low power transmissions are used to forward duta. However, evaluating the power level for the data packets is not sufficient. Reliable transmission of the data packets in a shared channel network requires a number of signalling messages at the Medium Access Channel (MAC) layer. The trmsmission power of the signalling messages is crucial for both reliable access and channel reuse. We propose a distributed range assignment algorithm (DRA), that can be applied to a power aware topology controi algorithm. to evaluate an appropriate pwer for the signalling messages. A worked example and a simulation based analysis of DRA, applied to a Distributed Relative Neighbourhood Graph is provided. Furthermore, a study on the impact of the transmission range, on the power usage and channel reuse is conducted. G. Srivastava http://works.bepress.com/pboustead/20 http://works.bepress.com/pboustead/20 Thu, 29 May 2008 09:48:12 PDT Power efficient topologies in an ad-hoc network can reduce battery usage and increase the lifetime of a network. Topology control algorithms including a local minimum spanning tree (LMST), distributed relative neighbourhood graph (DRNG) and K-Neigh graphs are computed by using the location or the distance information of the network nodes. Inaccurate distance or location information can lead to disconnected topology graphs. Furthermore, a distance based topology graph may not necessarily be connected due to wireless signal attenuation and propagation characteristics. In a realistic scenario, a small link distance may correspond to a large transmission power as the signal may need to transverse through obstacles. Hence it is imperative to include the environment characteristics when generating power efficient connected topology graphs. In this paper we identify the topology construction procedures that may lead to disconnected network topologies. We propose a new procedure that can work in conjunction with the neighbour discovery protocol to provide a better connected and power efficient network topology. Simulations indicate that the proposed algorithm provides a significant improvement in the connectivity of power base DRNG graphs. G. Srivastava http://works.bepress.com/pboustead/19 http://works.bepress.com/pboustead/19 Thu, 29 May 2008 09:48:09 PDT This paper develops a scalable system design for the creation, and delivery over the Internet, of a realistic voice communication service for crowded virtual spaces. Examples of crowded spaces include virtual market places or battlefields in online games. A realistic crowded audio scene including spatial rendering of the voices of surrounding avatars is impractical to deliver over the Internet in a peer-to-peer manner due to access bandwidth limitations and cost. A brute force server model, on the other hand, will face significant computational costs and scalability issues. This paper presents a novel server-based architecture for this service that performs simple operations in the servers (including weighted mixing of audio streams) to cope with access bandwidth restrictions of clients, and uses spatial audio rendering capabilities of the clients to reduce the computational load on the servers. This paper then examines the performance of two components of this architecture: angular clustering and grid summarization. The impact of two factors, namely a high density of avatars and realistic access bandwidth limitations, on the quality and accuracy of the audio scene is then evaluated using simulation results. P. Boustead http://works.bepress.com/pboustead/18 http://works.bepress.com/pboustead/18 Thu, 29 May 2008 09:48:06 PDT The current paradigm of content distribution networks (CDN) deployment requires heavy infrastructure investment since a large number of servers have to be deployed over a wide area. To overcome this difficulty, we propose a new paradigm where future CDNs are to be deployed over a leased server infrastructure. This paradigm shift changes the object placement problem in current CDNs to a combined object placement and resource provisioning problem. In this paper, we present a formulation of this problem, which is a combinatorial optimization, and prove its NP-hardness. We also present a heuristic solution to this problem based on Lagrangian relaxation and subgradient optimization, which also provides error bounds so that the accuracy of the obtained solution can be evaluated. The algorithm has been implemented and tested on a number of randomly generated network topologies, and results have shown promising solution accuracy. T. V. Nguyen http://works.bepress.com/pboustead/17 http://works.bepress.com/pboustead/17 Thu, 29 May 2008 09:48:04 PDT Topology control in a wireless ad-hoc network allows better spatial reuse of the wireless channel and network resources. The existing topology control algorithms tend to optimise network power usage by keeping the topology connected but do not take the network application requirements into account. Mission critical applications which require explicit end-to-end bandwidth and delay guarantees may not find enough resources in the network with the existing network topology. We have devised a topology control algorithm for mission critical applications (TCMCA) in wireless ad-hoc networks, which adapts the network topology to improve the available resources for a set of mission critical applications (high priority services) in a network. TCMCA is a source based algorithm where topology control decisions are made on distributed network knowledge. The performance of TCMCA is evaluated for a static wireless network and compared against algorithms such as connect, LINT and full power transmissions. We demonstrate that TCMCA shows better support for mission critical services for varying number of mission critical applications in the network. G. Srivastava