The error-recovery efficiency of packet-level forward-error-correction (PFEC) coding is sensitive to the loss-burst nature of mobile channels. Interleaving, which rearranges the packet transmission orders to scatter loss bursts, is commonly employed to help PFEC overcome this problem. Unfortunately, we still lack an efficient adaptation mechanism based on a tractable performance formulation that characterizes the tradeoffs between the interleaving and PFEC subject to network loss conditions (packet loss rate and loss correlation) and user requirements (buffer delay constraint and target packet loss rate). This paper addresses this issue by proposing an interleaving-first (IF) algorithm, in which we first pursue near-perfect interleaving based on a threshold mechanism, followed by optimizing PFEC throughput. Such an approach enables us to: (1) obtain a simplified performance model and (2) practice a simple systematic adaptation. The simulation confirms that the proposed scheme is capable of providing good quality sub-optimal solutions under typical mobile link conditions (i.e. the packet loss rate is ≤ 0.2, the mean burst length is ≤10 packets and the buffering delay is ≤256 packet times).

In mobile Peer-to-Peer networks (MP2P), packet losses are usually shown in bursts due to temporary peer failure or mobile channel disruption. This paper investigates how to apply Forward Error Correction (FEC) interleaving to overcome this issue and provide a guaranteed streaming service in MP2P, where a peer usually receives data from a set of heterogeneous peers through heterogeneous links. Some theories are developed for characterizing the problem, which help us to design a simple spiral packet-source allotment algorithm that implements near perfect FEC interleaving among sources with objective of maximizing throughput subject to an application-specific target packet loss rate. The simulation results confirm the superiority of the proposed scheme in aspects of combating burst loss and adaptability to peer/link heterogeneity for offering guaranteed streaming quality with high throughput under MP2P.

Wireless IP-based video broadcast suffers from heavy packet losses caused by multipath fading and interference variations in the wireless channels. This paper addresses this issue by proposing an application-layer error-resilience scheme, called the replicate multiple descriptor coding scheme (RMD). In principle, RMD extends the conventional multiple descriptor transmission strategy with two new features, the selective-frame-based replication and a time-shifted descriptor transmission. We show that with these two features, we are able to exploit the time diversity in a time-sharing channel to mitigate the damage impact and provide more efficient protection for the video. The simulation results confirm that when the packet loss rate is heavy (e.g., 15%–35%), RMD outperforms other schemes in terms of PSNR improved, while only requiring moderate data overheads.

Video streaming has been challenged by the best-effort service provided by the Internet. It is known that temporary traffic congestion may trigger the "traffic sawtooth" phenomenon of TCP and makes it inefficient in offering stable high-bandwidth transmission for video streaming. To address this issue, this paper proposes a segment-based flow control, called the adaptive TCP trunking (ATCP+). In principle, a video is transmitted in terms of a sequence of video segments. Each segment is then equally divided into several sub-segments and transmitted over distinct TCP connections in parallel. The size of sub-segment (or the number of TCP connections) is dynamically adaptive to the network conditions with the objective of stabilizing link throughput for a video stream. We implement ATCP+ with the HTTP 1.1 protocol. Practical experiments indicate that the proposed scheme is able to achieve a stable link throughput (e.g., 1.5Mbps) for a video stream under various network conditions.

Video multicast handoff in wireless networks is challenged by various impairments, such as high bit-error-rate, connection disruption and asynchronous stream concatenation. We tackle this problem by proposing a two-stage multicast approach, where a video stream is distributed to a group of access routers using a regular IP multicast. Within a local subnet, each access router further replicates and forwards the stream over a cluster of multicast channels by enforcing a staggered transmission schedule. A mobile host subscribes to all these channels and employs a buffer control to recover errors and to build up necessary backlogs. This approach gives us a scalable way to improve the error recovery rate and to smooth out potential stream gaps caused by the handoff. Some theoretical results concerning the buffer requirements for guaranteeing the seamless handoff are derived. The simulation results reveal that a cluster of three multicast channels along with a couple-second-video-clip buffer deployed in each access router and mobile host are capable of achieving satisfactory seamless video multicast handoff in most of the typical network environments.

This paper studies the buffer control issues that are related to supporting a seamless stream handoff in a WLAN environment that employs simulcast streaming. We analyze the impacts of the handoff disruption period and inter-AR synchronization variance. We also derive theoretical results that concern how a seamless stream handoff can be guaranteed in terms of the minimum buffer and backlog requirements. Based on the results, an MH-controlled-AR-pushed buffer control scheme is proposed and its cooperation with FMIPv6 is also discussed. The simulation results confirm the effectiveness of our scheme. In general, the deployment of a few-hundred-millisecond video-clip buffer in ARs and MHs is sufficient to achieve the seamless stream handoff most of the time, subject to fairly relaxed constraints of the inter-AR synchronization variance (ranging plusmn 150 ms) and handoff disruption period (around 50 ms).

This paper considers a scenario of broadcasting stream media over a mobile multicast networking environment, where a server employs the wired Internet multicast to replicate and distribute stream media to a set of access points (APs), from which mobile hosts (MHs) intercept the stream through wireless unicast links. An analytical framework is introduced for characterizing the traffic dependency between AP and MH by taking into account the issues impacting the seamless stream handoff (e.g., AP out-of-synch, handoff latency and path-setup delay). Based on it, a proactive hot-zone flow control scheme is proposed, which employs a simple threshold-based policy for regulating the traffic flow between AP and MH prior to handoff. The simulation results reveal that the proposed scheme can significantly reduce the playback hiccups while consuming limited buffer space, compared with the one without any special attention. Particularly, deployment of a few-second-video-length buffer in AP and MH is capable of achieving the seamless stream handoff most of the time subject to fairly loose constraints of the AP synch error (ranging /spl plusmn/3 sec) and handoff latency (around 1 second).

This paper investigates the implementation of video servers employing multi-zone disks. A multi-zone disk consists of a set of concentric disk zones with non-uniform storage capacities and data transfer rates. Superimposing a customized region layout on the top of such a disk is essential for facilitating the video server design and improving the disk resource utilization. This paper develops theories concerning the optimal region layouts for implementing the region-based data placement. Based on them, two efficient region layout algorithms are proposed, namely, the free-π scheme and fixed-π scheme. The former can achieve better resource utilization than the latter, but the latter requires a simpler timing control than the former. Extension of the fixed-π scheme to perform the disk striping in a disk array platform is also investigated. A spiral synchronization scheme is proposed to improve the system scalability by alleviating the disk bus congestion problem, caused by variable data transfer rates from individual disks. The analytical results reveal that the proposed schemes can effectively customize region layouts to achieve 85–99% resource utilization. By considering the region layout together with other design factors, an efficient configuration strategy is proposed for building a cost-effective server subject to the performance requirements of a target application.

A scalable quality-of-service (QoS) mechanism, termed harmonic DiffServ, is presented in this paper for supporting many-to-many heterogeneous-QoS multicasts in the DiffServ networks. In principle, it restricts data to be distributed using a limited number of multicast addresses (channels). In order to differentiate traffic with different QoS within a multicast channel, a small number of undefined DS codepoints (DSCP) are reserved for marking packets, and dynamic associations between DSCPs and service types are maintained in output ports of routers. We show that with proper arrangement on the usage of small numbers of multicast addresses and DSCPs, the forwarding information can be effectively shared among senders, which in turn significantly improves the system scalability.

The significant frame size variability exhibited in the compressed videos imposes a great challenge on network delivery. In this paper we propose an efficient flow control scheme, employed in the peer stations (i.e., servers and clients), for delivery of prestored compressed videos in a video-on-demand (VOD) system. This scheme resorts to an off-line analysis on the video frame sizes and server properties for figuring out the necessary buffer space and network bandwidth. The server platform of particular interest obeys a cycle-based data-block retrieval discipline, which is an essential technique to reduce the disk seek time for leveraging the disk throughput for supporting a large number of concurrent video accesses. Such a discipline is taken into account here to guarantee smooth delivery of variable-bit-rate videos. In run-time a server-driven control model is in use, where a server performs the primary flow control task, without relying on any feedback from clients. The scheme has been implemented in our prototype VOD system to support both unicast- and multicast communication paradigms under an RSVP-enabled network.

Exploitation of the statistical multiplexing gain in the delivery of compressed videos is essential for achieving efficient network resource utilization. This paper proposes an analytical framework based on a GOP-aware effective envelope for guaranteeing statistical quality-of-service (QoS) to an aggregate flow from a class of connections. The GOP-aware effective envelope intends to model more realistic traffic burst by taking the information of individual flow's launch time and group-of-picture (GOP) into account With this information, the QoS guarantee is applicable not only to the aggregate flow but also to individual flows. Through numerical experiments, we show that the significant improvement on multiplexing gain can be achieved, particularly, under those cases with short delay bounds and small-to-moderate aggregation scales.

Lightwave networks can be built by embedding virtual topologies over physical topologies. The optical passive star enables such embeddings easily. The Shuffle-net and the De-Bruijn graph are two popular virtual topologies proposed in the past for lightwave networks. Both however suffer from a lack of flexibility in scaling network sizes. We present a generalization of the De-Bruijn network which overcomes the limitation of the strict relationships between the network size parameters seen in the Shuffle-net and the De-Bruijn networks. A generalization for the Shuffle-net is also possible with this idea. We emphasize the support of time and wavelength division multiplexed media access protocols for such architectures and present several properties of the proposed network with respect to the same.

The paper examines the efficient layout of virtual paths (VPs) in an ATM network. The ATM network consists of ATM switches and their attached network end users, which may be gateways, routers, and hosts. The physical topology, the offered traffic, and call setup matrices of the network end users are assumed to be given. The problem is formulated as a flow-based optimization problem. A heuristic approach is presented which (i) establishes VPs according to physical network-specific and application-specific constraints and a cost function, (ii) provides multipaths between each source destination user pair to minimize the cell blocking probability and to increase network resilience, and (iii) uses a novel VP combining process which is guaranteed to always satisfy the switching constraints. Simulation results are presented for the proposed VP planning policy. Guidelines for the design of robust VP layouts and the efficient establishment of VCs are also presented.>

A dynamic hashing scheme based on extendible hashing is proposed whose directory can grow into a multilevel directory. The scheme is compared to the extendible hashing and the extendible hashing tree schemes. The simulation results reveal that the proposed scheme is superior than the other two with respect to directory space utilization, especially for files with nonuniform record distribution. This scheme can be easily extended to multikey file systems and also has good performance.>