Mikkel Christiansen, Kevin Jeffay, David Ott, and F. Donelson Smith, "Tunning RED for Web Traffic," in Proc. ACM/SIGCOMM'00, Stockholm, 2000.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....at this point, an thus the losses should occur at arriving in the buffer and duplicates would not account. To evaluate the previous issues, simulations were done to verify the relation between and the measured mean load in different intervals. The methodology applied in this work was based on [12]. In the simulations were used traffic sources with transfer sizes of 1, 5, 50, 500 and 5000 KBytes, and also sizes which obey the distributions shown in table I. The topology chosen to make the simulations is a kind of dumb bell and is presented in figure 1. The buffer size of the bottleneck ....

Mikkel Christiansen, Kevin Jeffay, David Ott, and F. Donelson Smith, "Tunning RED for Web Traffic," in Proc. ACM/SIGCOMM'00, Stockholm, 2000.

....congestion control, such as TCP, the service rate is not given by the ratio 1=L=C, since the transmission rate is regulated by the losses. In line with that, simulations were performed to identify the actual relation between average network load and . This approach follows the one introduced in [10]. The transfer sizes for long term connections was fixed at 256 KBytes. For short term connections, the transfer size follows a log normal distribution with mean 4827 bytes. Figure 2 illustrates the mentioned relationship. The functions that best fit the results were obtained by using the method ....

M. Christiansen, K. Jeffay, D. Ott, and F. D. Smith, "Tunning RED for Web Traffic," in Proc. ACM/SIGCOMM'00, (Stockholm), 2000.

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