Gary J. Anido

A cost comparison of distributed and centralized approaches to video-on-demand

With video-on-demand (VoD) regarded as one of the drivers for the deployment of broadband integrated service digital networks (B-ISDNs), an important issue is how to provide wide area VoD services most efficiently. A large contributor to the cost of a VoD system is storage. The nature of the service requires massive amounts of storage and bandwidth to be supplied from a video server located within the network. We examine the costs of storage in such servers and develop an efficient allocation scheme designed to minimize this cost. Using this scheme, centralized and distributed approaches to VoD are compared. We conclude that a distributed approach to storage costs no more than a centralized approach and offers considerable advantages in terms of bandwidth requirements and service quality.

Handover re-routing schemes for connection oriented services in mobile ATM networks

Wireless ATM (WATM) will be used to support broadband applications for future generation mobile services. To achieve this, the existing ATM protocol must be augmented with mobility management capabilities. One of the important components of mobility management is the ability of the network to seamlessly re-route existing connections to different parts of the network. To date, a number of schemes have been proposed for broadband wireless networks. These schemes can potentially be used in WATM networks. We compare the performance of these schemes in terms of their complexity, handover latency, communication disruption period and buffer and bandwidth requirements.

An analysis of multi-player network games traffic

Multi-player network games (also known as multi-player online games) are rapidly becoming a significant contributor to overall Internet traffic. However, the characteristics of the traffic generated by games are poorly understood, hence making it difficult to assess their impact (particularly on a large scale) on networks. This paper presents a study on the characteristics of multi-player games traffic, focusing on how the distributions of payload sizes and inter-arrival times vary with the number of players, thereby partially filling the aforementioned knowledge gap. We have found that these distributions are dependent on the game, but in some cases are independent of the number of players.

Multipath interconnection: a technique for reducing congestion within fast packet switching fabrics

The banyan interconnection is prone to internal link congestion, resulting in a blocking switch architecture. Several solutions that have been implemented to reduce the severity of link congestion offer packets a multiplicity of paths, which tend to increase packet delay variability and allow delivery of out-of-sequence packets. This, in turn, can lead to an increase in end-to-end protocol complexity, particularly in the case of real-time services. A solution called multipath interconnection is proposed to overcome this difficulty. Multiple (i.e., alternate) paths are provided and one is selected at call-setup time. Subsequent packets belonging to the call are constrained to follow the selected path. A number of path selection strategies are presented. >

Performability of disk-array-based video servers

Abstract. Video services are likely to dominate the traffic in future broadband networks. Most of these services will be provided by large- scale public-access video servers. Research to date has shown that disk arrays are a promising technology for providing the storage and throughput required to serve many independent video streams to a large customer population. Large disk arrays, however, are susceptible to disk failures which can greatly affect their reliability. In this paper, we discuss suitable redundancy mechanisms to increase the reliability of disk arrays and compare the performance of the RAID-3 and RAID-5 redundancy schemes. We use cost and performability analyses to rigorously compare the two schemes over a variety of conditions. Accurate cost models are developed and Markov reward models (with time-dependent reward structures) are developed and used to give insight into the tradeoffs between system cost and revenue earning potential. The paper concludes that for large-scale video servers, coarse-grained striping in a RAID-5 style of disk array is most cost effective.

Label switching and IP version 6

This paper discusses how the strict hierarchical address structure of IP version 6 and high levels of IP route aggregation will effect the operation of popular label switching techniques. We predict that the introduction of IP version 6 will cause significant problems with control driven label switching proposals such as MPLS and Tag Switching. We propose an alternate label switching solution that uses information in the IP version 6 address structure to classify and cut-through flows based on destination site address. Our simulation studies shown that over 96% of packets will be switched by ATM bypassing network layer forwarding using this technique.

Predictive call admission control for a disk-array-based video server

With video services featuring in future broadband applications, efficient video server architectures are required. Previous work has shown the efficacy of using large disk arrays for the provision of video to a substantial audience. Recent papers have proposed call- admission control schemes for video servers which utilize the statistics of compressed video and the characteristics of the disk drives themselves. To date, however, the issue of load-balancing has not been addressed. In this paper we supplement prior efforts with the development of a predictive call-admission control scheme which uses the largely deterministic nature of disk array operation to maintain a well-balanced load which in turn reduces the variation of the delay experienced by new requests. We provide detailed results comparing our scheme to others in the literature and show that the predictive scheme gives considerable improvement in terms of both admission delay and interactivity delay.

A concave link elimination (CLE) procedure and lower bound for concave topology, capacity and flow assignment network design problems

We examine the Concave Topology Capacity and Flow Assignment (TCFA) problem. Only two algorithms in the literature are appropriate for solving TCFA problems with concave link cost functions: Kleinrock and Gerlas Concave Branch Elimination (CBE) procedure [20,26] and a greedy link elimination procedure developed by Gersht [21]. However, neither works well in practice. The CBE procedure does not perform well in the context of strongly concave link cost functions. While Gershts algorithm performs well, its processing requirements are such that it is applicable for small network design problems only. We present a Concave Link Elimination (CLE) procedure, based on Gershts greedy link elimination procedure. Our algorithm is shown to perform at least as well as Gershts procedure and to be significantly faster than both the CBE and Gersht procedures. In addition, we formulate a lower bounding problem which we solve using a continuous branch-and-bound procedure to assess the quality of the design procedures.

Scalability and routing performance of label switching networks

Scalability is one of the major concerns associated with current label switching proposals. The aim of label switching is to provide an economical solution to the provision of high speed Internet IP networks. Previous research effort has concentrated on the issue of VC usage and its minimization. While VC usage is an important aspect of scalability the issues of routing and delay performance are equally significant. This paper considers in detail the routing and delay performance of various schemes. It is shown that the performance of control driven approaches such as tag switching depends upon the level of IP route aggregation in the network. Results also indicate that at high levels of aggregation, destination based approaches suffer delay and delay variance penalties over source/destination approaches such as IP switching.

Classification and comparison of schemes for providing scalable IP services over ATM networks

There are several approaches for the provision of scalable backbone IP over cell-based networks including multi-protocol label switching (MPLS) and multi-protocol over ATM (MPOA). This paper provides a classification that highlights the major differences between the various approaches. We classify approaches according to cut-through flow classification, path determination and forwarding granularity. This framework is used in conjunction with a comparative performance analysis to examine the strengths and weaknesses of popular approaches concentrating on MPLS and MPOA. We conclude that a new hybrid solution called destination-site label switching provides promising performance particularly in terms of flow classification and forwarding granularity.