Providing the personal live content delivery service using end user provided network resource is useful in many application situations due to its fast deployment. Together with service level agreement, the platform has to acquire sufficient network resources. Such resource acquisition can be realized by remuneration schemes for resource providers. From the perspective of service operators, selecting providers that effectively fulfilling SLA and efficiently reducing costs are their primary goals. We illustrate the feasibility of deploying Genetic Algorithms in searching resource providers economically. Our simulations showed that the effectiveness in satisfying the commitment on the resource provision stated in SLA and minimizing the related costs is highly related to its coding method and sensitive to the penalty factor. We find that the systematic choice of binary coding can improve the effectiveness of cost optimization in GA-based resource search.

In content delivery services, quality of service is greatly affected by the resources provided by the underlying overlay P2P. In order to maintain the level of resources acquired, soliciting peers' resources contributions by giving explicit remunerations is a feasible solution. From the service providers' perspectives, the remuneration should be minimized as long as the guaranteed services are fulfilled for all users. We consider a model defining the objective function to minimize this cost in service provision and a distributed and cost-efficient resource acquisition algorithm. Our results show that, in a network of the heterogeneous peers with limited local information, the service availability can be enhanced by appropriate resource acquisition schemes even in a heavily loaded network and the number of resource providers can be controlled by applying appropriate cost model.

In conventional video-on-demand systems, video data are stored in a video server for delivery to multiple receivers over a communications network. The video server's hardware limits the maximum storage capacity as well as the maximum number of video sessions that can concurrently be delivered. Clearly, these limits will eventually be exceeded by the growing need for better video quality and larger user population. This paper studies a parallel video server architecture that exploits server parallelism to achieve incremental scalability. First, unlike data partition and replication, the architecture employs data striping at the server level to achieve fine-grain load balancing across multiple servers. Second, a client-pull service model is employed to eliminate the need for interserver synchronization. Third, an admission-scheduling algorithm is proposed to further control the instantaneous load at each server so that linear scalability can be achieved. This paper analyzes the performance of the architecture by deriving bounds for server service delay, client buffer requirement, prefetch delay, and scheduling delay. These performance metrics and design tradeoffs are further evaluated using numerical examples. Our results show that the proposed parallel video server architecture can be linearly scaled up to more concurrent users simply by adding more servers and redistributing the video data among the servers.

In this paper, we propose the use of the Bidirectional Manhattan Street Network (BMSN) for interconnecting fixed size switching modules into a large scale ATM switch. Using BMSN for this purpose, each node now becomes an ATM switching module with its own input and output ports, and with four extension trunks connecting to four neighbouring modules. In this way, the switch size and capacity can increase by adding switching modules in a stepwise and planar manner, and there is no need for rewiring the I/O ports and the internal links. We extend BMSN into another architecture called SpiderNet, which connects eight instead of four neighbours. The average path length is reduced by having more direct paths. SpiderNet gives a better call blocking performance and cell loss probability, hence has a higher switch capacity. We show how routing paths can be established, and how cells are self-routed through the interconnected fabric. Cell loss and delay performance is also discussed.

Previously, we considered the design and implementation of a server array system for delivering video-on-demand service on a computer network. Data blocks are striped across an array of autonomous servers, and each client station contacts the servers one by one to retrieve blocks for constructing its own video stream. The server array approach has the benefits of (1) scalable storage capacity and throughput, and (2) load sharing across the servers. In this paper, we study fault-tolerance issues in a server array system and propose the notion of redundant array of inexpensive servers (RAIS), which is a server-level counterpart to redundant array of inexpensive disks (RAID). While some concepts from RAID may be applicable, we show that RAIS faces new challenges as failure detection and recovery are handled by network protocols, subjecting to error, loss, non-deterministic delay, and bandwidth limitation. Through an implementation of RAIS for delivering multimedia world-wide-web services, we show that (1) server-level fault tolerance can be achieved, (2) continuous playback of video and audio can be maintained in spite of server failures, and (3) graceful service degradation can be implemented using our object striping scheme.

Most video-on-demand (VOD) systems use a single, powerful server to deliver video streams to users. We consider a novel server array approach for delivering video services on networks. Such an approach has the benefits of (l) greater system capacity, as individual server has individual disk and network channel, (2) scalability, as more clients can be supported by adding more servers without data duplication, and (3) fault tolerance, as server-level fault-tolerant and fault-recovery schemes can be devised. We describe our experiences and results in the implementation of a server-array-based VOD system. Our system now has four P5-90 servers serving 40 PC-486 stations using a 10 Mbps Ethernet switch. The system can deliver 40 simultaneous and independent 1.2 Mbps, 30 fps, full TV size, MPEG-1 video streams. Our results demonstrate that with careful protocol design, Ethernet can deliver continuous video and audio services.

We consider a storage server architecture for multimedia information systems. While most other works on multimedia storage servers assume on-line disk storage, we consider a two-tier storage architecture with a robotic tape library as the vast near-line storage and on-line disks as the front-line storage. Magnetic tapes are cheaper, more robust, and have a larger capacity; hence they are more cost effective for large scale storage systems (e.g., video on demand (VOD) systems may store tens of thousands of videos). We study in detail the design issues of the tape subsystem and propose some novel tape scheduling algorithms which give faster response and require less disk buffering.>

Proposes a new fast packet switch architecture-pipeline banyan. It has a control plane and a number of parallel data planes which are of the same banyan topology. Packet headers are self-routed through the control plane to their destinations. As a result, they establish the corresponding routing paths in the data planes. The data planes do not need to do routing decisions, hence their complexity can be significantly reduced. Pipeline banyan can give a close to 100% maximum throughput and can deliver packets in a sequential order. Through analysis and simulation, the authors show that pipeline banyan has a better throughput and packet loss performance when compared with other banyan-type switch architectures.>

A new distributed protocol which supports concurrent message transmissions on different ring segments in a ring network is proposed. This protocol allows the destination station to remove the message body from the ring and to issue a new token for the succeeding stations to establish another transmission in the remaining ring segment. This protocol requires only one-bit latency at each station and supports variable size messages. We derived the maximum throughput of the pipeline ring and found it to be heavily dependent on the message size distribution. The maximum throughput of a single ring for exponential messages and fixed size messages are 1.23 and 1.68 respectively; while for the double ring, the per-ring throughput is 1.7 and 3.25 respectively. Due to analytical complexity, the delay performance is obtained by simulation. Two service disciplines are compared. It is found that the furthest within segment (FWS) discipline always performs better than the first come first serve (FCFS) discipline. The short message transmission scheme is introduced. Under bimodal traffic, it can significantly increase the ring efficiency.>

A ring protocol is proposed that allows voice and data traffic to coexist within the ring. This ring has the following distinct features: (1) it allows synchronous traffic such as voice and video to have a definite access to the channel within each packetization period (or frame); (2) it allows data messages to have a higher channel access priority provided that the synchronous traffic is not delayed by more than one frame; (3) it supports variable rate data circuits. Simulation results show that the data-message delay is much smaller than for other integrated services schemes. Urgent messages can be transmitted with a higher priority over voice. Since the voice packet delay is bounded within one packetization period, no time-stamping is needed and the voice loss can be completely avoided by reserving a sufficient number of slots. Continual speech reception is possible by synchronizing the speech regeneration process to the end of each frame. Since the ring is synchronized, gateway switching to external circuit-switched and packet-switched networks is very simple.>

The contention-based lookahead reservation (CLAR) protocol can provide fast circuit-switching services that are particularly advantageous for networks supporting integrated services. The delay and throughput performance for message transmission are obtained, and they agree closely with that obtained by simulation. The delay performance of CLAR is similar to that of the M-CSMA protocol for an M-channel network, but only CLAR can give a stable maximum throughput of (M-1)/M independent of the cable length. Moreover, CLAR requires only two sets of transceivers, while M-CSMA requires M. The lookahead reservation technique can provide 9% throughput increase for fixed-size messages and 19% for geometrically distributed messages.>