Hi, Hannes,
On 1/30/2012 4:24 AM, Hannes Tschofenig wrote:
I think we drifted away from the initial discussion.
We wanted to know whether
a) additional classes are needed, and
b) whether the currently proposed classes by Carsten need to be modified.
I actually think fewer are fine, as per below:
I heard folks saying
a) yes - we need a class 0 for very small devices, and
b) the current classes need to be modified (although I haven't really seen a
convincing proposal).
I gave a definition with two classes:
Limited IP <<100 packets
supports ARP/ND, IP, and UDP
can support DHCP, TCP, and HTTP in limited forms
typically does not support dynamic routing or complete
versions of DHCP, TCP, HTTP, or other protocols
Full IP >>100 packets
not limited in what it *can* support
(might be limited by choice in what it *does* support)
IMO, the middle ground is fuzzy anyway.
Joe
I believe it also makes sense to briefly describe what we believe can be done
with devices of the various classes.
For example, a device from class 0 can already run IP, DNS, DHCP and a UDP/TCP
based client side application with a pre-shared secret key based security
mechanism.
Ciao
Hannes
_______________________________________________
Lwip mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lwip
_______________________________________________
Lwip mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lwip
