No abstract

No abstract

In this paper, we propose a scheme for estimating the proportion of the incoming traffic that is not responding to congestion at a router. The idea of the proposed scheme is that if the observed queue length and packet drop probability do not match with the predicted results from the TCP model, then the error must come from the non-responsive traffic; it can then be used for estimating non-responsive traffic. The proposed scheme utilizes queue length history, packet drop history, expected TCP and queue dynamics to estimate the proportion. We show that the proposed scheme is effective over a wide range of traffic scenarios through simulations.

No abstract

In this paper, we propose a scheme for estimating the proportion of the incoming traffic that is not responding to congestion at a router. The idea of the proposed scheme is that if the observed queue length and packet drop probability do not match with the predicted results from the TCP model, then the error must come from the non-responsive traffic; it can then be used for estimating non-responsive traffic. The proposed scheme utilizes queue length history, packet drop history, expected TCP and queue dynamics to estimate the proportion. We show that the proposed scheme is effective over a wide range of traffic scenarios through simulations.