Instead of the increuse-by-one decrease-to-half strategy used in TCP for congestion window adjustment, we consider the general case such that the increase value and decrease ratio are parameters. That is, in the congestion avoidance state, the window size is increased by a per window of packets acknowledged and it is decreased to p of the current value when there is congestion indication. We refer to this window adjustment strategy as general additive increase multiplicative decrease (GAIMD). We present the (mean) sending rate of a GAIMD flow as a function of a, p, loss rate, mean round-trip time, mean timeout value, and the number ofpackets acknowledged by each ACK. We conducted extensive experiments to validate this sending rate formula. We found the formula to be quite accurate for a loss rate of up to 20%. We also present in this paper a simple relationship between a and p f o r a GAIMD flow to be TCP-friendly, that is, for the GAIMD flow to have approximately the same sending rate as a TCPflow under the same path conditions. We present results from simulations in which TCP-friend111 GAIMDflows (a = 0.31, p = 718) compete for bandwidth with TCP RenoJlows and with TCP SACKflows, on a DropTail link as well as on a RED link. We found that the GAIMDflows were highly TCP-friendly. Furthermore, with / 3 at 7/8 instead of 1/2, these GAIMD flows have reduced rate fluctuations compared to TCP flows.

No abstract

Abstract-We investigate the fairness, smoothness, responsiveness, and aggressiveness of TCP and three representative TCP-friendly congestion control protocols: GAIMD, TFRC, and TEAR. The properties are evaluated both analytically and via simulation by studying protocol responses to three network environment changes. The first environment change is the inherent fluctuations in a stationary network environment. Under this scenario, we consider three types of sending rate variations: smoothness, short-term fairness, and long-term fairness. For a stationary environment, we observe that smoothness and fairness are positively correlated. We derive an analytical expression for the sending rate coefficient of variation for each of the four protocols. These analytical results match well with experimental results. The other two environment changes we study are a step increase of network congestion and a step increase of available bandwidth. Protocol responses to these changes reflect their responsiveness and aggressiveness, respectively.

We consider the problem of mining association rules over interval data (that is, ordered data for which the separation between data points has meaning). We show that the measures of what rules are most important (also called rule interest ) that are used for mining nominal and ordinal data do not capture the semantics of interval data. In the presence of interval data, support and confidence are no longer intuitive measures of the interest of a rule. We propose a new definition of interest for association rules that takes into account the semantics of interval data. We developed an algorithm for mining association rules under the new definition and overview our experience using the algorithm on large real-life datasets.

No abstract