Journal of High Speed Networks - Volume 6, issue 1

Authors: Chen, Wen-Tsuen | Hwang, Wu-Yuin

Article Type: Research Article

Abstract: In this paper, cell loss priority schemes in the ATM switches are studied. Many schemes had been proposed to manage buffer access for priority classes. However, these schemes suffered from poor buffer utilization or prohibitive implementation complexity. Therefore, we propose a simple priority scheme called Separate logic Queues with Complete buffer Sharing (SQCS). In the SQCS scheme, buffers are shared by priority classes and two service strategies, First-Come-First-Service (FCFS) and High-Priority First (HPF), are employed to serve priority classes. The performance of our proposed scheme is evaluated and compared with those of other priority schemes. The results indicate that our …proposed scheme has a better performance and low implementation complexity. Hence, the SQCS scheme is an effective priority scheme for the ATM switching system. Show more

Keywords: ATM, QOS, buffer management

DOI: 10.3233/JHS-1997-6101

Citation: Journal of High Speed Networks, vol. 6, no. 1, pp. 1-14, 1997

Authors: Wu, Jieh-Chian | Willner, Alan E. | Silvester, John A.

Article Type: Research Article

Abstract: The use of low-speed tunable devices may lead to throughput and delay penalties for packet-switched WDM star networks. When the tuning latency is significant, conventional cost-effective solutions result in significant throughput and delay penalties. Our approach: (i) combines and unifies existing approaches to improve performance by pipelining multiple tunable filters per node and making a TDM schedule with maximum time gap for reception; (ii) introduces ‘superpackets’ (multiple packets with the same source and destination), thereby relaxing the impact of large tuning latency. Since the superpackets increase throughput at the cost of increased delay, we determine the minimum delay, under a …constraint of no throughput penalty. A discrete-time queueing model of the protocol shows that a system with two tunable filters per node has low delay and is wavelength-efficient for uniform traffic. Show more

DOI: 10.3233/JHS-1997-6102

Citation: Journal of High Speed Networks, vol. 6, no. 1, pp. 15-31, 1997

Authors: Anastasi, G. | Lenzini, L.

Article Type: Research Article

Abstract: This paper analyzes in depth the asynchronous part of the MetaRing MAC protocol by considering a network configuration (hereafter scenario) in which the slot reuse capability is used to the minimum extent possible. The MetaRing MAC protocol behavior is first analyzed in asymptotic conditions (i.e., when each station is trying to seize all the medium capacity), then in underload conditions (i.e., when the offered load is lower than the maximum achievable aggregate throughput), and finally in overload conditions (i.e., when the offered load is slightly higher than the maximum achievable aggregate throughput). Our results show that the MetaRing is unfair …and that the unfairness depends, in particular, upon the workload parameters, in addition to the K and L protocol parameter values. Our analysis includes medium capacities in the range 150 Mbps-2.4 Gbps. Show more

DOI: 10.3233/JHS-1997-6103

Citation: Journal of High Speed Networks, vol. 6, no. 1, pp. 33-52, 1997

Authors: Sholander, Peter E. | Owen, Henry L.

Article Type: Research Article

Abstract: Pointer processors in Synchronous Optical Networks (SONET)/Synchronous Digital Hierarchy Networks (SDH) may be placed into three categories: fixed threshold pointer processors, adaptive threshold modulation pointer processors and uniform pointer processors. Fixed threshold pointer processors produce gapped pointer sequences during SONET/SDH holdover mode when one or more node's clock has lost its synchronization reference. These gapped pointer sequences cause mapping wander in the underlying PDH payloads. This wander has motivated the development of uniform pointer processors which reduce/remove the gaps caused by the STM-n, AU and TU overhead bytes. However, uniform pointer processors cannot remove the additional gapped pointer effects caused …by the VC overhead bytes. This paper reviews previous mathematical analysis and simulation results for SDH networks which use fixed threshold pointer processors to transport E1 PDH payloads in VC-12. That analysis is then extended to SDH networks which transpOlt DSl in VC-11. Next, a uniform pointer processor algorithm is proposed for the case where VC-12 transports an E1 rate PDH payload. Finally, a lower bound on the Maximum Time Interval Error (MTIE) that can be achieved with a uniform pointer processor is calculated. Show more

Keywords: Synchronous Digital Hierarchy, Synchronous Optical Network, uniform pointer processor, network synchronization, network jitter minimization

DOI: 10.3233/JHS-1997-6104

Citation: Journal of High Speed Networks, vol. 6, no. 1, pp. 53-65, 1997

Authors: Bissonauth, Ashvanee | Kamal, Ahmed E.

Article Type: Research Article

Abstract: In this paper, we introduce a new protocol, which we refer to as the R2 -Net protocol. The objective of this proposal is to introduce an access protocol that is based on allowing active nodes to share the bandwidth in a round robin manner. By following such an approach, it is shown that the protocol does not suffer from the problems inherent in the DQDB protocol. The implementation of the round robin principle requires nodes to keep track of the number of active downstream nodes. In addition, and in order to overcome the adverse effects of the nonzero propagation delay, …the total number of segments that must be transmitted by downstream nodes is also recorded. The protocol has several attractive features such as fairness in bandwidth distribution among nodes, a speedy convergence to such a distribution, a low access delay, and more importantly, insensitivity of such measures to node activity patterns. A simulation study has shown the satisfaction of such properties. Since priority is an important feature of every protocol, we also show how the R2 -Net protocol can be generalized to implement priority. Show more

Keywords: Computer-communication networks, network protocols, local networks, metropolitan area networks, simulation and modeling

DOI: 10.3233/JHS-1997-6105

Citation: Journal of High Speed Networks, vol. 6, no. 1, pp. 67-79, 1997