Hi,

On 21/05/2020 20:30, Alexandre Petrescu wrote:
>> For AX.25 version 2.0, the
>>    maximum frame size expected is 330 bytes and implementations MUST be
>>    prepared to handle frames of this size.  Higher frame sizes can be
>>    negotiated by AX.25 version 2.2 and so this is a minimum requirement
>>    and not a limit.
> 
> It is ok.   For IPv6, the minimum MTU (minimum Maximum Transmission
> Unit) is 1280 bytes.
> For IPv4, it is not worth mentioning.  One, including myself, would go
> as far as suggesting to remove the IPv4 keyword from the draft altogether.

I'm happy to reword this as long as the semantic meaning does not
change, however, removing IPv4 would effectively make this process
meaningless. 100% of the current deployment (that I'm aware of) uses
IPv4. Any hope of transitioning towards IPv6 will depend on a transition
step, which requires the co-existence of the two protocol versions for a
time.

For IPv4, the minimum size that a host must accept is 576, which is
greater than the minimum MTU. 330 is less than 576 but I don't think it
is redundant to specify the size of the AX.25 frame that you end up with
after decapsulation.

> A packet dump, or an illustration diagram, showing an AX.25 carried in
> IPv6 would be helpful.

This would be a very simple diagram I guess:

+-----------------+
|   IP Header     |
+-----------------+
|  AX.25 Header   |
+-----------------+
|  AX.25 Payload  |
+-----------------+

Or, including security considerations recommendations:

+-----------------+
|   IP Header     |
+-----------------+
|   ESP Header    |
+-----------------+
|  AX.25 Header   |
+-----------------+
|  AX.25 Payload  |
+-----------------+

> AX.25 is related to X.25 but remark: X.25 might carry IP packets,
> whereas IP packets might carry AX.25.  If I understand correctly.

The relationship between AX.25 and X.25 is really of historic interest.
It is possible to put AX.25 inside IP, and it's possible to put IP
inside AX.25. There are several schemes used for encapsulating IP in
AX.25 however none of them are defined in IETF documents. As I implement
them, I may try to write up more drafts. Let's see how this one goes first.

Thanks,
Iain.

-- 
https://hambsd.org/

_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area

Reply via email to