[Olsr-dev] [RFC] Weighted HNA plugin

[Clauz]

But collisions might happen also with "normal" HNAs and addresses. That
is a matter of addressing policy. For example a specific subnet could be
allocated to fake nodes, to recognize them easily.

Clauz


On 07/04/2012 08:02 PM, Henning Rogge wrote:
> The typical thing would be an address collision. Either with another
> "special HNA address" or a main address.
> 
> Henning Rogge
> 
> On Wed, Jul 4, 2012 at 2:27 PM, Clauz <(spam-protected)> wrote:
>> On 07/03/2012 03:39 PM, Henning Rogge wrote:
>>> On 07/03/2012 04:26 PM, Clauz wrote:
>>>> Hi!
>>>> I have developed an idea by Saverio of a weighted HNA plug-in. The aim
>>>> is to have a fixed/set ETX for announced HNA networks, but preserving
>>>> compatibility with nodes without the plug-in.
>>>>
>>>> So we introduced a fake OLSR node announcing the HNA and linked with the
>>>> wanted ETX metric.
>>>>
>>>> This is achieved through "internal packet injection" code.
>>>>
>>>> https://github.com/ninuxorg/olsrd/blob/whna_plugin/lib/whna
>>>> https://github.com/ninuxorg/olsrd/blob/whna_plugin/lib/whna/README_WHNA
>>>
>>> I know they added this into the OLSRv2 draft, because they forgot to add
>>> it into the packet format in OLSRv1.
>>>
>>> But I don't know if a hack like this is really a good idea, especially
>>> because we generate some "bogus" rule to the IP address of the faked node.
>>
>> Do you have any specific "disaster scenario" in mind?
>>
>> Clauz
>>
>>
>> --
>> Olsr-dev mailing list
>> (spam-protected)
>> https://lists.olsr.org/mailman/listinfo/olsr-dev
> 
> 
>