Differentiated end-to-end Internet services using a weighted proportional fair sharing TCP

Abstract

In this document we study the application of weighted proportional fairness to data flows in the Internet. We let the users set the weights of their connections in order to maximise the utility they get from the network. When combined with a pricing scheme where connections are billed by weight and time, such a system is known to maximise the total utility of the network. Our study case is a national Web cache server connected to long distance links. We propose two ways of weighting TCP connections by manipulating some parameters of the protocol and present results from simulations and prototypes. We finally discuss how proportional fairness could be used to implement an Internet with differentiated services.

An Extension of the TCP Steady-State Throughput Equation for Parallel Flows and Its Application in MulTFRC

“…As stated in [9] and as our comparison shows, MulTCP performs reasonably well with values up to . This is shown in Fig.…”

“…Flow does not experience loss in every TD-period, so the TD-periods in which flow experiences loss are a subset ( , ) of TD-periods of the cumulative flow. If is the window size of flow at the end of the th TD-period, and is the number of rounds from the end of till the end of , during rounds the window size of flow increases by (as in [1]) and we have (9) The number of packets sent in a TD-period is the number of packets sent between two loss events of the cumulative flow. For the th TD-period, this includes packets sent in the last round of the th TD-period, starting from the th packet till the end of the window ( packets, as shown in Fig.…”

“…Every rounds, the window size of the cumulative flow is increased by , and we have (10) Assuming that and are mutually independent sequences of i.i.d. random variables (as in [1]), from (9) we have (11) Taking into account the assumption that a loss occurs identically distributed over all flows we can say that and are equal for all flows, and from now on we denote them by and . From (2) and (5), we have (12) We assume that at the end of a TD-period , flows experiencing loss in that TDP have the window size , and other flows that experience loss in the previous loss event have the window size , etc.…”

“…We tested MulTCP [9] in the same simulation setup as CP using an update that is available under "contributed code" via the main ns-2 Web page to make it work with the most recent version of ns-2. As stated in [9] and as our comparison shows, MulTCP performs reasonably well with values up to .…”

Policing congestion response in an internetwork using re-feedback

“…Re-feedback ensures that a policer at the network ingress 6 From Eqn (5) an attacker can reduce dropping probability by increasing variance, e.g. by alternating honest & zero packets.…”