Journal of High Speed Networks

In this paper we present an implementation‐based comparison of Measurement‐based Admission Control algorithms. Through the use of a special purpose environment, a performance and behaviour comparison is made. The results of this paper illustrate the independence of traffic from admission control behaviour in the homogeneous traffic environment. While illustrating the impact the admission decision will make upon heterogenous traffic systems. These results highlight the importance of estimators being robust to statistical variation in measurements and offering calibrated controls. While based upon a comparison of measurement‐based admission control algorithms — these conclusions are applicable to any application of measurement‐based estimation.

This paper presents a new robust TCP congestion‐recovery scheme to (1) handle bursty packet losses while preserving the self‐clocking capability; (2) detect a TCP connection's new equilibrium during congestion recovery, thus improving both link utilization and effective throughput; and (3) make the TCP behavior during congestion recovery very close to that during congestion avoidance, thus “extending” the performance model for congestion avoidance to that for TCP loss recovery. Moreover, the new recovery scheme requires only a slight modification to the sender side of TCP implementation, thus making it widely deployable. The performance of the proposed scheme is evaluated for scenarios with many TCP flows under the drop‐tail and RED gateways in the presence of bursty packet losses. The evaluation results show that the new scheme achieves at least as much performance improvements as TCP SACK and consistently outperforms TCP New‐Reno. Furthermore, its steady‐state TCP behavior is close to the ideal TCP congestion behavior. Since the proposed scheme does not require selective acknowledgments nor receiver modifications, its implementation is simpler than SACK, and only the servers in the Internet need to be modified while keeping intact millions of clients scattered in the Internet.

We investigate the behavior of a new response strategy to TCP Explicit Congestion Notification (ECN). The new strategy is more aggressive in the short term, but preserves TCP long term behavior – without modifying the router ECN marking rate. A more aggressive short term behavior gives incentives for hosts to become ECN‐compliant. ECN serves as an early warning sign in this case. Our analysis demonstrates the effectiveness of the new TCP ECN behavior. Simulation results with short/long lived FTP, UDP and HTTP connections, multiple bottleneck configurations, and various TCP flavors and parameters, demonstrate higher throughput and reduced oscillations with the new response strategy.

We consider the problem of efficient grooming of low‐speed connections onto high‐capacity lightpaths in WDM mesh networks. Maximizing the throughput and minimizing the network cost are the two most important objectives of the traffic grooming problem. It may not be possible to achieve both the goals simultaneously. We present an efficient heuristic algorithm to maximize throughput without compromising on the network cost. Our algorithm decides the placement of the transceivers at the nodes instead of operating on a fixed number of transceivers. Multihop routing and rerouting techniques have been used to achieve higher throughput. We use a two stage ILP approach to enable comparison of our algorithm with the optimum. We also present simulation results to contrast the performance of our approach with two recently proposed heuristic algorithms.

Path selection under multiple additive QoS constraints in high‐speed networks is an NP‐complete problem. Most of the algorithms proposed for QoS routing suffer from either excessive computational complexities or low performance. In this paper, we propose two practical heuristic algorithms that are based on iterative shortest‐path computations with a dynamic cost function. The cost function comprises a combination of several QoS constraints and a single optimization metric. A simulation program evaluates the performance of the algorithms. The simulation results indicate that our algorithms give very high performance and fast running time.