[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ND-proxy applicability and loop-prevention



Pekka,
      Maybe I missed it, but how can we decrement the Hop Count
for loop prevention when receiving nodes are using the Hop
Count=255 check on reception?

Regards,
Brian

Pekka Savola wrote:

> Btw -- one aspect was mentioned by Mohan P. on v6ops list: whether the
> ND-proxy could decrement TTL instead of keeping it the same.  I don't
> think whether this would affect this discussion (i.e., whether such a
> proxy would be considerably better in this respect) has been
> considered.
> 
> On Thu, 25 Mar 2004, Erik Nordmark wrote:
> 
>>>The same helpdesk she calls when she encounters a weird problem in her 
>>>network connectivity, or in her PC.  Most likely you ;-)
>>>
>>>(This is a much more generic problem, not one specific to this 
>>>scenario, obviously.)
>>
>>But in this particular case you seem to be arguing that plug&pray
>>is sufficient while I argue that we should aim for plug&play;
>>I think the futuristic goal is that wiring together network devices
>>shouldn't be more complex than plugging in electrical appliances.
> 
> 
> For the deployment I have in mind, plug&pray and plug&play are pretty 
> much equivalent.  There are certainly other things, labeled plug&play 
> which are much more brittle than this :)
> 
> 
>>>You are making assumption that those boxes would also be acting as
>>>routers (in the ND-proxy mode) by default, right?  I don't, and I
>>>don't think doing that would make a lot of sense.
>>
>>No.
>>I'm only making the same assumption that underlies ndproxy as well as the 
>>zerouter discussion; there will be L2s that do not support IEEE 802 bridging.
>>If you disagree with this assumption we don't need ndproxy or zerouter
>>for the home networking case - IEEE 802 bridging has already solved the
>>problem.
> 
> 
> OK -- maybe you're thinking of this in more generic terms, like, every 
> VCR or equivalent having its own (more or less) internal media for 
> which IP connectivity would be desirable.  And that such media would 
> not be currently IEEE bridgeable. (And when you combine this to a 
> scenario when such a "VCR++" has WLAN uplink to the other devices at 
> home, you might end up in a mess.  With wired connectivity, you'd be 
> OK.)
> 
> On the other hand, I have been looking at the scenario where an 
> explicit set of "routers" (or a home PC or whatever) would be proxying 
> for the nodes connecting to that box.
> 
> We certainly seem to have an indication that the latter is important.  
> I do not have personal knowledge if the former would be.  It'd be 
> interesting to know.
> 

--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------