| M. Csenger, "Early ATM Users Lose Data", Communications Week, May 16, 1994. |
.... the performance of ATM for besteffort data traffic (exceptions are [C91, SC93, HKM93] 2 ) Since ATM does not provide Media Access Control, it has been a concern that the throughput will be low if an ATM network experiences congestion; in fact there is already practical evidence to this effect [C94]. This paper uses simulation studies to investigate the throughput behavior of TCP over ATM for best effort traffic when there is network congestion. First, we consider the throughput performance of TCP over ATM without any additional ATM AAL level 3 congestion 1 The ATM Forum is an ....
M. Csenger, "Early ATM Users Lose Data", Communications Week, May 16, 1994.
.... problem of the relevance of fractal traffic in practice, see [4, 5, 102, 105, 115 117, 121, 122, 138, 146, 184, 199,258, 318 320, 328,335, 356] In this context, see also the discussions in [117, 318, 413] related to some practical experience with the first generation ATM buffers reported in [79]. While there is considerable scope for future research in the area of queueing models with long range dependent inputs, queueing in the presence of heavy tailed service time distributions (and, in general, independent arrivals) is relatively well understood; e.g. see [1, 2, 56, 72, 213, 323, ....
M. Csenger. Early ATM users lose data. Communications Week, May 16 1994.
....1995 9 traffic models are significant for buffer sizing; all other factors being equal, the heavier asymptotics would imply larger buffer requirements. Many first generation ATM switching systems had under dimensioned buffers, leading to unacceptable cell losses when subjected to real traffic [6]; while there are several reasons why first generation switches had under dimensioned buffers, an awareness of self similar traffic modeling would have sensitized designers to the potential for this problem. Another engineering issue of interest is the setting of statistical multiplexing gains. ....
M. Csenger, "Early ATM Users Lose Data", Communications Week, May 16, 1994.
....cooperation. ATM for best effort data traffic (exceptions are [C91, SC93, C93, HKM93] Since ATM does not provide Media Access Control, it has been a concern that the throughput will be low if an ATM network experiences congestion; in fact there is already practical evidence to this effect [C94]. This paper uses simulation studies to investigate the throughput behavior of TCP over ATM for best effort traffic when there is network congestion. First, we consider the throughput performance of TCP over ATM without any additional ATM AAL level 2 congestion control mechanisms, such as ....
M. Csenger, "Early ATM Users Lose Data", Communications Week, May 16, 1994.
.... As networks move towards Broadband ISDN (B ISDN) based on ATM (Asynchronous Transfer Mode) technology for transport [8, 35] there is considerable interest in devising techniques for enhancing or insuring Quality of Service (QoS) via adequate control [27] 29] 30] 31] A recent article [7] in Communications Week describes users frustrations with some ATM products inability to handle congestion and bursty traffic over Wide Area Networks (WANs) Sources in the article attribute this poor performance to inadequate buffer allocation in switches and to a lack of intelligent buffer ....
Csenger, M. "Early ATM Users Lose Data," Communications Week, May 1994.
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