Sharing some thoughts … as already pointed on this list, we will more likely 
need two routing protocols,
one of constraint objects (e.g. sensors, actuators, ..), one for the other 
devices, this is perfectly fine (and
we know that some devices will not need routing at all, fine too). With regards 
to Smart objects, since the
IETF has already specified a routing protocol (RPL in the ROLL WG), do we now 
want to "Tweak"/"change"
/Add yet a third one (a simplified LS or routing protocol such as OLSR) ? Since 
it was mentioned that we 
should select a well established and stabilized routing protocol, it is worth 
pointing out that although RPL 
is fairly new, in the world of smart objects, this is the protocol that will 
rapidly get the largest adoption and
experience in the field, see the number of alliances that adopted it 
(Zigbee/IP, Wave2M, …) 

On Oct 4, 2011, at 3:50 PM, Fred Baker wrote:

> 
> On Oct 4, 2011, at 12:50 AM, Lorenzo Colitti wrote:
> 
>> For example, if there were a robust open-source version of IS-IS that didn't 
>> take up too much RAM and flash, you could define TLVs to implement zeroconf 
>> and you could get to something that works relatively quickly.
> 
> RFC 1453 specified a default OSPF configuration that my (ACC) Router came up 
> with by default. The only configuration necessary was to supply IPv4 
> prefixes/addresses to the interfaces and to turn OSPF on. In the context of a 
> home, I suspect it would be pretty close. Fundamentals: default values for 
> most attributes are as specified in the OSPF RFC, the metric is the duration 
> of a bit in 10 ns units (which in LAN networks becomes a hop count), and all 
> interfaces are in area zero. Change things to your heart's content, but the 
> default configuration is one that actually works for most purposes in a small 
> network. I'll entertain arguments that this doesn't address every possible 
> configuration, but for the networks we're discussing, I'll argue that its 
> "close enough".

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to