Re: [Autoconf] Call for comments to a new AUTOCONF charter proposal.

[Mark Townsley <[e-mail]>]

On 7/3/10 9:57 AM, Teco Boot wrote:
> Easy to invent the automatic mechanism:
> Only need an address? Use ND.
> Need more? use ND for getting address, find central server, get more.

OK, as described that's still one mechanism per type of information. So,
where does the DNS server address go ;-)

- Mark

> 
> Teco.
> 
> 
> Op 2 jul 2010, om 20:59 heeft Mark Townsley het volgende geschreven:
> 
>>
>> My kneejerk reaction to this is that walking in with the goal of having
>> more than one way to autoconfigure a manet is a bad idea.
>>
>> If we end up with two ways to autoconfigure, then we will have to invent
>> an automatic mechanism on top to choose which autoconfiguration
>> mechanism to use.  That doesn't help anyone. In absence of knowledgeable
>> human configuration, hard choices that narrow functional options
>> typically far outweigh the potential benefits of one option vs. the
>> other. So, even if you can prove that A is better than B, B is still
>> better than A+B.
>>
>> Let's strive for making a choice, at least within the MANET domain.
>>
>> - Mark
>>
>>
>> On 7/2/10 3:21 PM, Holtzer, A.C.G. (Arjen) wrote:
>>> Hello autoconfers,
>>>
>>> I support this "two-case"-approach, Christopher mentions: so
>>> standardizing one centralized and one decentralized solution (or one
>>> stateful and one stateless solution, just like in the current IPv6
>>> standards). I agree that the solution should make use of existing
>>> protocols as much as possible (e.g. DHCP, ND, ...), but my choice would
>>> be not to state in the charter that DHCP must be used in all solutions
>>> coming out of the WG.
>>>
>>> draft-bernardos-manet-autoconf-survey-05 shows there are already many
>>> proposals existing, making it a good starting point for going into
>>> solution space. Actually even more than just a starting point since many
>>> of the proposals have already been around for a while. So I support this
>>> doc.
>>>
>>> Best regards,
>>> Arjen
>>>
>>>>
>>>> If Charlie can find a few like-minded people to work on that, 
>>>> why not add this as a parallel activity? The rationale of why 
>>>> two cases should be straightforward to make, they are almost 
>>>> chalk and cheese in e.g. centralised versus non-centralised. 
>>>> This is actually added safety to the group producing 
>>>> something, as if one succeeds and the other fails, that's still good.
>>>>
>>>>
>>> This e-mail and its contents are subject to the DISCLAIMER at http://www.tno.nl/disclaimer/email.html
>>>
>>> _______________________________________________
>>> Autoconf mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/autoconf
>>>
>>
>> _______________________________________________
>> Autoconf mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/autoconf
> 
>