[c-nsp] ME3400 FHRP - GLBP with IPv4 and IPv6

[Gerald Krause]

Am 02.12.2013 16:29, schrieb Lukas Tribus:
> Hi,
> 
>> Just FYI - we've got the answer from TAC now that the ME3400G won't
>> support an FHRP with IPv4 and IPv6 on the *same interface*. It seems to
>> be a hardware limitation of this series.
> 
> Does this apply only to ME3400 non-E, or to the E series as well?
> 
> ME3400 non-E is EOL and EOS next month (that could be standard excuse nr1):
> http://www.cisco.com/en/US/prod/collateral/switches/ps6568/ps6580/end_of_life_notice_c51-726108.html

Our call was for the old non-E version. We should had tried it much much
earlier...

> 
> 
> 
>> Call my imagination somewhat limited, but I have a truly hard time
>> understanding how you could build a hardware that would not be able
>> to do that. This is basically a pure control-plane functionality,
>> plus "handle a virtual MAC address" - and if you use the same vmac for
>> IPv4+IPv6, I can't see how the forwarding hardware would need to know
>> that there is IPv6 involved as well.
> 
> Agreed, just punt the packet to the CPU, its not like using one destination
> mac with multiple applications would be a first: 01:00:0c:cc:cc:cc is used for
> CDP/VTP/DTP/PAGP/UDLD - and that even includes parsing an LLC header to
> differentiate the application; here we just need to differentiate based on the
> EtherType.
> 
> 
> Probably the "business case" is not big enough to fix this bug, and declaring
> a WONTFIX for such a severe IPv6 limitation in 2013 would not be very wise
> from a PR perspective, so they made up the hardware excuse.

...because now they can use this excuse to bridge the time gap until
end-of-everything and no one will do anything more on this platform.


Gerald