cc_newreno4

Usage

Overview

Details about the algorithm can be found in RFC5681.

Options

The cc_newreno module supports a number of socket options under TCP_CCALGOOPT (refer to tcp(4) and mod_cc(9) for details) which can be set with setsockopt(2) and tested with getsockopt(2). The cc_newreno socket options use this structure defined in <sys/netinet/cc/cc_newreno.h>:

struct cc_newreno_opts {
	int name;
	uint32_t val;
}

CC_NEWRENO_BETA

Multiplicative window decrease factor, specified as a percentage, applied to the congestion window in response to a congestion signal per: cwnd = (cwnd * CC_NEWRENO_BETA) / 100. Default is 50.

CC_NEWRENO_BETA_ECN

Multiplicative window decrease factor, specified as a percentage, applied to the congestion window in response to an ECN congestion signal when net.inet.tcp.cc.abe=1 per: cwnd = (cwnd * CC_NEWRENO_BETA_ECN) / 100. Default is 80.

Note that currently the only way to enable hystart++ is to enable it via socket option. When enabling it a value of 1 will enable precise internet-draft (version 4) behavior (subject to any MIB variable settings), other setting (2 and 3) are experimental.

Note that hystart++ requires the TCP stack be able to call to the congestion controller with both the newround function as well as the rttsample function. Currently the only TCP stacks that provide this feedback to the congestion controller is rack.

MIB Variables

The algorithm exposes these variables in the net.inet.tcp.cc.newreno branch of the sysctl(3) MIB:

beta

Multiplicative window decrease factor, specified as a percentage, applied to the congestion window in response to a congestion signal per: cwnd = (cwnd * beta) / 100. Default is 50.

beta_ecn

Multiplicative window decrease factor, specified as a percentage, applied to the congestion window in response to an ECN congestion signal when net.inet.tcp.cc.abe=1 per: cwnd = (cwnd * beta_ecn) / 100. Default is 80.

Acknowledgements

Development and testing of this software were made possible in part by grants from the FreeBSD Foundation and Cisco University Research Program Fund at Community Foundation Silicon Valley.