From 95c0790a88711ba1c3821df200e10c6841c3a0a9 Mon Sep 17 00:00:00 2001 From: Adam Ierymenko Date: Thu, 8 Aug 2013 12:55:01 -0400 Subject: [PATCH] Back off a little on multicast propagation depth. Eventually this will be a network parameter. --- node/Constants.hpp | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/node/Constants.hpp b/node/Constants.hpp index f1a36c65b..49322030f 100644 --- a/node/Constants.hpp +++ b/node/Constants.hpp @@ -226,9 +226,16 @@ error_no_ZT_ARCH_defined; * the sum of BREADTH^i where I is from 1 to DEPTH. This ignores the effect * of the rate limiting algorithm or bloom filter collisions. * - * 7 results in a max of 21844 recipients for a given multicast. + * 5 results in a max of 1364 recipients for a given multicast. With a limit + * of 50 bytes/sec (average) for multicast, this results in a worst case of + * around 68kb/sec of multicast traffic. FYI the average multicast traffic + * from a Mac seems to be about ~25bytes/sec. Windows measurements are TBD. + * Linux is quieter than Mac. + * + * This are eventually going to become per-network tunable parameters, along + * with per-network peer multicast rate limits. */ -#define ZT_MULTICAST_PROPAGATION_DEPTH 7 +#define ZT_MULTICAST_PROPAGATION_DEPTH 5 /** * Length of ring buffer history of recent multicast packets