So, I read that there is no limit to the number of IPv6 addresses that can be assigned to an interface.
Doesn't that mean that someone could then assign ALL the IPv6 addresses to an interface and exhaust the supposedly inexhaustable address space in one go?
:awesome:
You're close, actually this leads to tcam exhaustion for multicast state as it has to join a new multicast group per address.
I would also assume the IPv6 stack would have a limit on the number it can support for the hardware. Not sure though.
There was a bug about a year ago where you could crash a Windows box by sending 100 router advertisements with different networks. It would setup an address for itself on each one, and eventually fall over.
-Otanx
This is fun.
OK, so for my plan to work, apparently I need a MAJOR memory upgrade on a router... increase the number of tcam entries possible... make similar adjustments in the IPv6 stack...
Quote from: deanwebb on January 21, 2017, 08:38:49 AM
This is fun.
OK, so for my plan to work, apparently I need a MAJOR memory upgrade on a router... increase the number of tcam entries possible... make similar adjustments in the IPv6 stack...
Or stand up a linux based router within a VM say Cumulus or quagga and go to town.
Now to find a BGP peer to work with...
Quote from: Otanx on January 20, 2017, 07:43:39 PM
There was a bug about a year ago where you could crash a Windows box by sending 100 router advertisements with different networks. It would setup an address for itself on each one, and eventually fall over.
-Otanx
Here it is: https://nmap.org/nsedoc/scripts/ipv6-ra-flood.html
It will knock out everything in the VLAN, unless the vendors have already implemented a fix.
edit: it seems they implemented a fix in Win8.1
Great, so we can have a Windows host run a batch file and get those IPs on there!
:lol: