W. Leland, "Window Based Congestion Management in Broadband ATM Networks : The Performance of Three Access-Control Policies," IEEE Globecom'89 Conf.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... delays of messages and cells, if the offered connectionless traffic is shaped before entering the ATM network [1, 3] It was also learnt from the above studies that access delays form the most significant part in the end to end delays, which confirms the simulation results obtained by Leland [31], as well. Traffic shaping means a preventive traffic control action. According to ATM Forum User Network Interface Specification Version 3.0, recommendations are as follows: ffl When used in a private ATM switch, traffic shaping is a mechanism that alters the traffic characteristics of a stream ....

W. E. Leland. Window-based congestion management in broadband ATM networks: The performance of three access-control policies. In GLOBECOM'89, Dallas. TX., 1989.

....function that attempts to avoid or control congestion and limit its undesirable effects. Approaches to congestion control include reserving resources in advance of their use [7, 10, 34] discarding excessive cells when overload occurs [9, 30] and shaping the traffic before it enters the network [3, 29, 38]. Most conventional congestion control schemes rely on end to end exchange of control messages in order to regulate traffic flow. The control messages (sometimes with additional congestion information added by the intermediate nodes) are used as feedback by the source node, which then adjusts its ....

....only if its required 12 performance could be met. Each connection s traffic flow is enforced to remain within its allocated bandwidth. For preventive congestion control schemes for high speed networks, many researchers suggest that rate based approaches are preferable to window based approaches [3, 29, 38, 46, 47, 48]. Since most network resources, such as transmission links and switch processors, are rate limited (i.e. they cannot sustain packets arriving at a rate faster than their capacity) transmission rate can provide a proper match between user demands and the resources available. Another argument for ....

[Article contains additional citation context not shown here]

W. Leland, "Window-Based Congestion Management in Broadband ATM Networks: The Performance of Three Access Control Policies", Proceedings of INFOCOMM 1989, Ottawa, Canada, pp. 49.7.1-49.7.7.

....large networks. Recently, Golestani [7] has proposed a stop and go service policy that allows rate enforcement at intermediate nodes. Third, with dynamic rate based controls there is a possibility of significant increase in packet queue when the total input rate is close to the capacity. Leland [8] shows that rate based controls result in a higher total delay (including delay at the source queue, as well as in the network) and a higher loss rate than window based controls at loads in excess of 60 of the capacity. Depending upon the feedback and control delay, the queue lengths may increase ....

W. E. Leland, "Window-based Congestion Management in Broadband ATM Networks: The Performance of Three Access-Control Policies," Proc. Globecom'89, Dallas, TX, November 27-30, 1989, Vol. 3, pp. 1794-1800.

.... at the user network interface or at a data source, by buffering and injecting cells into the network at a slower speed [5] The smoothing function may be guided either by the desire to conform to a given traffic rate, end to end acknowledgment, or end to end acknowledgment with congestion feedback [6]. The invocation of both admission control and a smoothing function is not necessarily contingent on the performance prediction. The smoothing function applies primarily to the admittance of new cells from the external user, and not cells that are forwarded to a switching node from other switching ....

W.E. Leland, "Window-Based Congestion Management in Broadband ATM Networks: The Performance of Three Access-Control Policies," GLOBECOM'89, Vol. 3, 1989, pp. 1794-1800.

....another value j, with j i. Our work differs from [6] in that we directly characterize those periods of time in which arriving customers are lost, rather than characterizing loss as being quasi stationary during periods of times during which the number of active sources remains constant. Leland [7] mentions, but does not elaborate on measuring consecutive losses per connection in an ATM simulation experiment. Woodruff and Kositpaiboon [8] mention the desirability of specifying the probability and duration of periods of high cell loss rates. Ferrandiz and Lazar [9] investigate the ....

W. E. Leland, "Window-based congestion management in broadband ATM networks: the performance of three access-control policies," in Proceedings of the Conference on Global Communications (GLOBECOM), (Dallas, Texas), pp. 1794--1800, IEEE, Nov. 1989.

No context found.

W. Leland, "Window Based Congestion Management in Broadband ATM Networks : The Performance of Three Access-Control Policies," IEEE Globecom'89 Conf.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC