[rbridge] Max Network size / ARP servers
touch at ISI.EDU
Wed May 11 15:44:17 PDT 2005
-----BEGIN PGP SIGNED MESSAGE-----
Guillermo Ibáñez wrote:
> would consider big a network (single broadcast domain) with 25.000
> -100.000 hosts. But this could fall short.
> As stated in the paper, the ARP cache policy in hosts (Windows) is as
> follows: unused entries in last two minutes expire, the refreshed ones
> are allowed a maximum of 10 minutes, then a new ARP request will be
> sent. Measurements, (see below at the end of mail) are based on the
> current caching at endhosts, so the caching effect is already included.
> Regarding snooping, I agree that snooping of broadcast responses at
> proxy-ARPs will reduce the load.
We agree that this is a problem for a sufficiently large network. As
I've already noted, there are two things RBridges are not uniquely
trying to solve:
a) increased size of L2 subnets
b) insecurity of ARP
Solutions to either should work fine in an RBridge scenario, but are not
part of the prerequisites of the RBridge architecture.
Although an RBridge may encourage large subnets - larger than are
currently typical - so do large L2 switches. There are solutions in that
space to reduce broadcasts (IGMP snooping, proxy ARP, etc.) that might
apply just fine here, but aren't worth (IMO) mentioning explicitly.
As you noted, there are plenty of challenges with proxy ARP - hashing,
load balancing, fault tolerance, etc. But all those solutions will
benefit all L2 subnet systems, and are not specific to RBridges.
>>>Some measurements on ARP load are available at:
> I did not refer to this paper to support the solution proposed, but only
> as reference for the ARP measurements provided. Results were: 89 ARPs
> per second on average in a 2.456 hosts network, with peak load of 1150
> ARPs/second. For a network with 25.000 hosts (ten times size), this
> means 900 ARPs /second on average and 11.500 ARPs packets (peak) to
> process on every host.
Just because there are 10x as many hosts doesn't mean there will be 10x
as many ARPs, notably because there won't typically be 10x as many
routers to the rest of the Internet.
There are also other economies of scale possible in an RBridge -
broadcasts can be more efficient than in a spanning tree because there
can be multiple broadcast trees inside the RBridge campus.
Overall, as said before, I don't think this is an issue specific to
RBridges. Does anyone else??
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
More information about the rbridge