[olsr-dev] auto ip address assignment

[Abhishek Misra]

well if hashing is proper then what is the probability of collision of
these hash values.  

n Wed, Apr 5, 2006, Jens Nachtigall <(spam-protected)> said:

> --nextPart1201449.ej62VUQJhn
> Content-Type: text/plain;
>   charset="iso-8859-15"
> Content-Transfer-Encoding: quoted-printable
> Content-Disposition: inline
> 
>> > I think in the freifunk firmware, which most
>> > meshes use, also the MPRs are switched off, i.e. every node behaves lin=
> ke
>> > an
>> > MPR. So the hierarchie is flat. This is because, the MPR idea did not
>> > work out.
>>
>> I didn't know that... it sounds quite frustrating. MPR seemed to be a good
>> idea...
> 
> Yes, MPRs seem to be one of these ideas which work out nicely in theory, bu=
> t=20
> once the code is there and you start experimenting in a real world mesh, it=
> =20
> did not work out that nicely. But for your ip auto stuff this is actually=20
> good news, because the duplicate ip address detection becomes much easier,=
> =20
> because every node (not only MPRs) sends link state (i.e. topology) message=
> s,=20
> so you get the seq. number plus its originators ip address from every node.=
> =20
> so the ideas of pacman are easier to implement. Anyway, if you have a good=
> =20
> auto ip generation (ie hashing from the mac addr) the detection will be use=
> d=20
> rarely. But i would not wait for the best of all ideas, just start somethin=
> g,=20
> we test it in reality, and then it goes forward in small steps.
> 
> 
> jens
> 
> --nextPart1201449.ej62VUQJhn
> Content-Type: application/pgp-signature
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.1 (GNU/Linux)
> 
> iD8DBQBEM36RAzJOgw63+4oRAi3wAJ4yCRe3nGUOLSx/7cPHonNB0KgVHQCfQOkn
> MO/pXkgb6ACHt6zvqArKHTM=
> =9Gbu
> -----END PGP SIGNATURE-----
> 
> --nextPart1201449.ej62VUQJhn--
> 
> 

--