Hengheng Xie

A Multipath Video Streaming Solution for Vehicular Networks with Link Disjoint and Node-disjoint

The high quality video streaming in vehicular networks is an urgent topic to provide services on safety and infotainment on the road. To provide high quality video streaming, Forward Error Correction (FEC) is one of the most popular approaches to ensure the needed quality for video streaming by generating duplicated packets. However, several factors might cause problems to FEC in a VANET, like the limited resources of wireless networks, a highly dynamic topology and the large amount of data in video streaming. The duplicated packets might exceed the network capacity. Therefore, in this work, the retransmission mechanism is used to ensure the transmissions, rather than the FEC. The main problem of the retransmission mechanism is the delay. A multi-path solution based on a disjoint algorithm is proposed to reduce the interference and contention, leading to a higher transmission rate and an acceptable delay. In this solution, only I-frames are transmitted through the TCP protocol, and the inter-frames are transmitted through the UDP protocol. To improve the delay of TCP transmissions, a TCP-ETX algorithm is integrated to select the suitable path for TCP transmissions. Simulations are conducted and several results are presented by comparing them to other protocols. Based on the simulation results, the designed multi-path solution protocol provides a higher video quality with a reasonable delay than that of the other protocols.

MERVS: A Novel Multichannel Error Recovery Video Streaming Protocol for Vehicle Ad Hoc Networks

Accompanying the increasing interest in vehicle ad hoc networks (VANETs), there is a request for high-quality and real-time video streaming on a VANET, for safety and infotainment applications. Video streaming on a VANET faces extra issues, in comparison with the video streaming on a mobile ad hoc network (MANET), such as the highly dynamic topology. However, there are also benefits to VANETs, such as large buffer and battery capacity, predictable motion of vehicles, and powerful central and graphic processing units (CPU and GPU, respectively). However, the high packet loss ratio of a VANET is a critical issue for high-quality video streaming. In this paper, we propose an error recovery process for high-quality and real-time video streaming in a VANET, which is call multichannel error recovery video streaming (MERVS). MERVS transmits the video through two different channels: a reliable channel and an unreliable channel. Because of the importance of the intraframes (I-frames) in terms of video quality, I-frames will be transmitted through the reliable channel. The interframes will be transmitted through the unreliable channel because of the limited resource of the reliable channel. The priority queue, quick start, and scalable reliable channel (SRC) techniques are also integrated to improve the delay of MERVS. Based on the conducted simulation results, MERVS can provide higher quality video streaming compared with forward error correction (FEC) with similar time delay compared with the real-time transport protocol/user datagram protocol (RTP/UDP) in a VANET.

A novel cross layer TCP optimization protocol over wireless networks by Markov Decision Process

Transmission Control Protocol (TCP) is well-known for its poor performance over wireless networks. The main reason is the unstable condition of wireless networks. Previous attempts to improve TCP performance have been centered on techniques such as queue management, congestion window adjustment, delayed acknowledgement, and etc., because lower layers are hidden in the layered Open Systems Interconnection (OSI) model. With cross layer techniques, more precise data can be retrieved from lower layers for network condition estimation. In this paper, we analyze and model the factors that affect the TCP throughput in the lower layers. Consequently, to solve the TCP optimization problem, we formulate TCP performance as a Markov Decision Process (MDP). An extensive set of simulation experiments has been performed on NS-2 in order to illustrate that TCP performance can be enhanced by cooperatively adjusting lower layer factors.

A novel collision probability based adaptive contention windows adjustment for QoS fairness on ad hoc wireless networks

It is crucial to achieve Quality of Service (QoS) on IEEE802.11 in order to provide stable and reliable communication for real time and multimedia applications. Most of the recent QoS techniques for Ad Hoc Networks rely on basic QoS classifications such as Enhanced Distributed Channel Access (EDCA) and Hybrid Coordination Function Channel Access (HCCA) with stationary backoff range for specific services, without considering overall resource consumption and the external correlation among terminals. In this paper, a novel contention window adjustment QoS scheme, as a special case of QoS classification, is proposed for achieving QoS-fairness on Ad Hoc Networks. It applies dynamic automatic QoS assignment with low complexity by considering the restriction among various terminal requirements. Throughput estimation with a proportional relation between throughput and backoff parameters is discussed based on a Markov chain model of a saturated Ad Hoc Networks. We measure the performance of different QoS algorithms theoretically and verify improvements of our algorithm on QoS-fairness with detailed simulations.

Design of A QoS-Aware Service Composition and Management System in Peer-to-Peer Network Aided by DEVS

QoS-aware service management and composition has become an interesting research topic with the rapid development of service oriented architecture (SOA). Differently with Web-services based systems, the emergence of peer-to-peer (P2P) based distributed network technology brings more challenge to the QoS-aware service management and composition. In this paper, we propose our design of a QoS-aware hierarchical service composition and management system in a context of JXTA-enabled P2P network. We conducted a comparison experiment of our design with commonly used flat-based service composition and management, and found that our design outperforms the flat-based one in terms of a higher success rate for satisfying the users QoS requirement. Furthermore, we used a RT-DEVS model based formal approach to validate our design, and believe that it can be a promising technology in aiding the design of an efficient QoS-aware service composition and management system.

An Efficient Time Management Scheme for Large-Scale Distributed Simulation Based on JXTA Peer-to-Peer Network

As an emergence technology, P2P is spreading to distributed simulation area, and many distributed simulation frameworks have used P2P as the middleware to interconnect their existing single processors simulators to form distributed environments for simulation execution. In terms of simulation time management, most existing tools use a middleware layer to implement and support time management in a heterogeneous networking environment, and therefore, it is generally not easy to migrate a single processors simulation to multi-processors in these frameworks. In this paper, we present a P2P based distributed simulation time management based upon JXTA API and Service Oriented Architecture (SOA), and we focus our discussion on how we implement the time management as a JXTA peer and a JXTA group service. Our time management is actually a native distributed message passing management framework, and does not rely on any middleware layer. Furthermore, we evaluate the performance of our implementations using a local Linux cluster. This work will establish a solid foundation for the more advanced distributed simulation services that have been proposed in our project [1].

TCP-ETX: A cross layer path metric for TCP optimization in wireless networks

Transmission Control Protocol (TCP) suffers degradation in wireless networks because, among other factors, low quality links tend to interrup a TCP connection. Most of the routing protocols use the number of hops as path metric without considering the link condition. Expected Transmission Count (ETX) calculates the estimated number of transmissions to successfully transmit a packet, but it fails to calculate the correct number of transmissions for a TCP segment. In addition, the overhead of ETX is very high because it needs to probe all destinations to calculate the link loss probability. In this paper, we propose a novel cross layer path metric, TCP-ETX, which calculates the accurate number of transmissions for TCP segments and minimizes the protocol overhead. To evaluate our proposal, we conduct simulations and compare our solution to existing protocols, which shows that our protocol can offer a significant improvement on the TCP performance.

A QoS-aware load balancing algorithm for P2P based large-scale distributed virtual environment

Distributed Virtual Environment (DVE) has attracted much attentions in recent years due to the rapid advances in the areas of E-learning, Internet gaming, human-computer interfaces, and etc. Meanwhile, in recent years, the emergence of Service Oriented Architecture (SOA) provides a more flexible and inter-operable framework for building more advanced DVE based applications. In such a background, service composition and management becomes a major concern, especially when considering an efficient QoS-aware load balancing scheme in the system. In this paper, we propose and implement a QoS-aware adaptive load-balance algorithm that aims to provide dynamic and adaptive load-balanced service composition and management in a JXTA-enabled P2P network based DVE. We compare our algorithm with other commonly used schema in P2P network, and find that ours outperforms the others in many aspects by a series of real-time distributed simulation experiments in a dynamic service-oriented P2P network.

TCP-CC: cross-layer TCP pacing protocol by contention control on wireless networks

Transmission Control Protocol (TCP) performs poorly over wireless networks. Some research indicates that the TCP congestion control mechanism may cause burstiness in the traffic flow. Numerous TCP segments are delivered simultaneously, while an acknowledgement of a retransmission is successfully received. Burstiness leads to a highly contentious network, which markedly increases the probability of packet loss on wireless networks. TCP pacing is a possible solutions for TCP burstiness on multi-hop networks. In this algorithm, TCP segment transmissions are distributed over the whole Round Trip Time. Most pacing protocols attempt to insert a delay interval into the TCP transmissions. However, there is a similar pacing algorithm in IEEE802.11, known as the contention window mechanism. In this paper, we first measure and analyze the way that the contention window size affects TCP throughput in different scenarios. We propose a cross-layer TCP pacing protocol by contention control in the MAC layer, called TCP Contention Control (TCP-CC). It adjusts the lower bound of the contention window in order to optimize the overall TCP throughput in both one-hop and multi-hop topology. Finally, comparative simulations are conducted in order to verify the improvements of our protocol on both TCP Reno and TCP Vegas.

A novel cross layer TCP pacing protocol for multi-hop wireless networks

Some research indicates that the TCP congestion control mechanism may cause the burstiness of traffic flow. A group of TCP segments are delivered simultaneously, while an acknowledgement (ACK) of a retransmission is successfully received. This burstiness leads to a network with high level contention , which extremely increases the probability of packet loss on wireless networks. TCP pacing is one of the solutions for TCP burstiness on multi-hop network. The idea is to spread the TCP segment transmissions over the whole Round Trip Time (RTT). Most of the pacing protocols attempt to insert a delay interval into the TCP segment flows or ACK flows, without noticing a Cross Layer pacing algorithm on IEEE802.11 with deferring and backoff mechanism. In this paper, we measure and analyze the way contention window size (CW) affects TCP throughput in different scenarios; and, we propose a cross layer TCP pacing protocol by controlling the lower bound of contention window for the optimization of the overall TCP throughput on multi-hop topology. Finally, a comparison of simulations is presented in order to verify the improvement achieved by our protocol.