Re: [radext] #192 (bigger-packets): unclear meaning of "configured categories" of dynamically discovered servers

[Stefan Winter <[e-mail]>]

Hi,

> I think the particular configuration you quote above might be a good fit
> for a system that uses both Moonshot and eduroam.
> Especially if you make it a bit finer grained by allowing  each
> consortuim/set of CAs/whatever you want to describe it as for DNS have
> its own configuration.

The "whatever you want" makes it difficult to prescribe how exactly a
classification of servers should be done - and was the reason why I
considered the wording unclear.

> I think I need something like this as a minimum so I'd like to have a
> related SHOULd in the spec.
> I'd be happy to get input on what approaches would give me this but work
> for you.

Okay... We both agree that this parameter needs to be configurable. None
of us really knows how "categories" should be defined exactly (nor do we
know what will be useful classification criteria in the field).

I think this is something actual implementations - and their customer
demand - can and will sort out themselves.

The important bit is that it must be configurable in principle IMHO; and
that more than one global default is something to consider for an
implementation. So how about this as a text proposal:

OLD:

If dynamic discovery mechanisms are supported, configuration SHOULD be
provided for the maximum size of clients and servers in each dynamic
discovery category.

NEW:
If dynamic discovery mechanisms are supported, the default maximum
packet sizes for sending packets SHOULD be configurable. Implementations
MAY choose to allow configuration of different defaults in different
parts of their dynamic discovery configuration.

How about that?

Greetings,

Stefan Winter


-- 
Stefan WINTER
Ingenieur de Recherche
Fondation RESTENA - Réseau Téléinformatique de l'Education Nationale et
de la Recherche
2, avenue de l'Université
L-4365 Esch-sur-Alzette

Tel: [phone]
Fax: [phone]

PGP key updated to 4096 Bit RSA - I will encrypt all mails if the
recipient's key is known to me

http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xC0DE6A358A39DC66