V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Commun. Rev., 29(3), July 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....curve I in Fig. 3. For simplicity we ignore this additional delay. To actively drop packets with the exact probability obtained above is difficult to realize due to the computation complexity. Instead we propose to use a simpler scheme similar to the gentle variant of Random Early Detection (RED) [7] to approximate the expiration rate. For the values of RED parameters like w q and max p , we set them as suggested in [8] In Fig. 3 we present two piece wise linear AQM functions II and III to approximate the Gaussian curve. It is understandable that the better the approximation is, the better ....

V. Rosolen, O. Bonaventure and G. Leduc, "A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic", ACM Computer Communication Review, vol. 29, no. 3, Jul. 1999.

....delay = 50 ms. The loss probability grows linearly when the number of connections is small. D( q) 0 if q min th 1 if q 2 max th q min th max th min th max p if q 2 [min th ; max th ] q max th max th (1 max p ) max p otherwise As shown by Rosolen et al.[39, 40], the gentle option makes the RED much more robust to the setting of the parameters max th and max p . Therefore, we turn it on for all simulations in this paper. Moreover, in this paper, we only consider RED without support for ECN. 5.2. RED Configurations RED has four control parameters: ....

V. Rosolen, O. Bonaventure and G. Leduc. A RED Discard Strategy for ATM Networks and Its Performance Evaluation with TCP/IP Traffic. ACM Computer Communication Review, July 1999.

....min th to max th . However, when the average queue size is greater than max th , RED drops all incoming packets until the average queue size drops below max th . Studies show that RED can improve throughput and fairness over drop tail queue management while maintaining a low average queuing delay [6, 10]. However, this benefit is only achieved for well configured RED under some traffic loads, specifically when the average queue length does not significantly oscillate and stays under max th . Other researchers conclude that RED is too complicated to configure, and show that end to end ....

....performances significantly. Although general RED configuration guidelines that work well for a large set of traffic load have been proposed [4] RED configuration difficulties will exist where Internet traffic varies. As a fix to the above problem, the gentle modification to RED was proposed [10], which replaces the packet drop behavior when the average queue size is over max th as shown in Figure 2 (a) Instead of setting the drop probability to 1 after the average queue size goes over max th , gentle RED linearly increases the drop probability from max p to 1 as average queue size grows ....

V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Communication Review, Jul 1999.

....to . However, when the average queue size is greater than , RED drops all incoming packets until the average queue size drops below . Studies show that RED can improve throughput and fairness over drop tail queue management while maintaining a low average queuing delay [6, 10]. However, this benefit is only achieved for well configured RED under some traffic loads, specifically when the average queue length does not significantly oscillate and stays under . Other researchers conclude that RED is too complicated to configure, and show that end to end ....

....performances significantly. Although general RED configuration guidelines that work well for a large set of traffic load have been proposed [4] RED configuration difficulties will exist where Internet traffic varies. As a fix to the above problem, the gentle modification to RED was proposed [10], which replaces the packet drop behavior when the average queue size is over as shown in Figure 2 (a) Instead of setting the drop probability to 1 after the average queue size goes over , gentle RED linearly increases the drop probability from to 1 as average queue size ....

V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Communication Review, Jul 1999.

....th q Gammamin th max th Gamma min th Delta max p if q 2 [min th ; max th ] q Gammamax th max th Delta (1 Gamma max p ) max p otherwise q B 2 max th max th min th max p 1 D(q) Fig. 1. Drop function of RED with the gentle modification. As shown by Rosolen et al. 32] [33], the gentle option makes the RED much more robust to the setting of the parameters max th and max p . Therefore, we turn it on for all simulations in this paper. RED has four control parameters: min th , max th , max p , and w q . How to properly configure these parameters has been the ....

V. Rosolen, O. Bonaventure and G. Leduc, "A RED Discard Strategy for ATM Networks and Its Performance Evaluation with TCP/IP Traffic," ACM Computer Communication Review, Jul. 1999.

No context found.

V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Commun. Rev., 29(3), July 1999.

No context found.

Rosolen V., Bonaventure O. and Leduc G, A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic, 1999. Available from: http://citeseer.nj.nec.com/rosolen99red.html.

No context found.

V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Commun. Rev., 29(3), July 1999.

No context found.

V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Commun. Rev., 29(3), July 1999.

No context found.

V. Rosolen, O. Bonaventure, and G. Leduc. A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic. ACM Computer Commun. Rev., 29(3), July 1999.

....per VC scheduling and Random Early Detection (RED) implementation (figure 4) the minimum guaranteed bandwidth is provided by the utilization of a per VC scheduler. The buffer acceptance mechanism relies on the ATM version of RED described in [18] Other variants of RED could have been used [19] [20]. 9 This mechanism It attempts to accomplish fair sharing of the buffer resources by maintaining accounting information for each per VC queue. This buffer acceptance algorithm is configured by specifying two global thresholds ( F badc and VeRf6adc , not shown in figure 4 ) and two thresholds ....

V. Rosolen, O. Bonaventure, and G. Leduc, "A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic," ACM Computer Communication Review, July 1999.

No context found.

V. Rosolen, O. Bonaventure, and G. Leduc, "A red discard strategy for atm networks and its performance evaluation with tcp/ip traffic," to appear in Computer Communication Review, Vol. 29, No. 3 , July, 1999.

No context found.

J V. Rosolen, O. Bonaventure and G. Leduc,"A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic", ACM Computer Communication Review, vol. 29, no. 3, Jul. 1999.

No context found.

O. Bonaventure, V. Rosolen, and G. Leduc, "A RED discard strategy for ATM networks and its performance evaluation with TCP/IP traffic," ACM Computer Communication Review, vol. 29, no. 3, July 1999.

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