[Olsr-dev] Ninux GSoC 2012, ideas on olsrd

Teco Boot (spam-protected)
Tue Mar 20 06:57:44 CET 2012


Op 19 mrt. 2012, om 17:56 heeft Henning Rogge het volgende geschreven:

> mDNS was always a variant of BMF for the special case of Zeroconf. So
> its not unreasonable to think about how both should be joined.
> 
> p2pd has a different approach to multicast forwarding (sparse vs.
> dense multicast).
The main difference is that p2pd has filter parameters so packets are classified for flooding.
mDNS has a hard-coded filter. A bit ugly. 

Teco

> 
> Henning Rogge
> 
> On Mon, Mar 19, 2012 at 17:04, Teco Boot <(spam-protected)> wrote:
>> 
>> Op 19 mrt. 2012, om 12:48 heeft ZioPRoTo (Saverio Proto) het volgende geschreven:
>> 
>>> Hello,
>>> 
>>> 1) fix mDNS plugin
>>> we explain this point in detail here:
>>> http://wiki.ninux.org/GSoCIdeas2012#Improve_the_OLSRd_mDNS_plug-in
>> 
>> 
>> We fixed some issues in mDNS and released this as p2pd. p2pd can easily act as mDNS replacement. I can think of a merge with bmf also, where some mcast packets are transfered as OLSR messages (low overhead) and some are send with (encapsulated) IP packets. We need a filter syntax for classification (ignore / OLSR-message /  IP packet).
>> 
>> I use both bmf and p2pd, where broadcasts are taken by p2p2 and multicasts are taken by bmf. bmf has no filter itself, I use iptables on LAN port. Is a Q&D trick.
>> 
>> Some code in p2pd for de-duplication was not completed. But why are we replicating bmf features?
>> 
>> This leads to my advise: get rid of mDNS and p2pd. Let's improve bmf. Maybe there is a code-size problem. It  is a few 10kB larger than mDNS an p2pd. And will increase when enhanced. I don't have problems with that.
>> 
>> Teco
>> --
>> Olsr-dev mailing list
>> (spam-protected)
>> https://lists.olsr.org/mailman/listinfo/olsr-dev
> 
> 
> 
> -- 
> Steven Hawkings about cosmic inflation: "An increase of billions of
> billions of percent in a tiny fraction of a second. Of course, that
> was before the present government."





More information about the Olsr-dev mailing list