[IPsec] More comments on draft-mglt-ipsecme-diet-esp-07

[Robert Moskowitz <[e-mail]>]

Continuing at sec 6.1:

Skipping 6.2 for now, as it will not be used for current use case (I 
realize I may have one for Manned Aircraft).

Good til 7.2, then skipping 7.2 and 7.3 for now.

I like 7.4 in that UDP gets compressed to zero bytes.  And the way you 
have constructed diet-esp to include transport, a separate SCHC rule for 
transport is not needed.  Now if the payload is CoAP, then things will 
be different.  Per the rfc 8824.

Skip 7.5 and 7.6

Sec 11:

    Security Parameter Index (SPI):
       Until Diet-ESP is not deployed outside the scope of IoT and small
       devices,


r/ not / /

?

What is that not doing there?

    Sequence Number (SN):  If incremented for each ESP packet, the SN may
       leak some information like the amount of transmitted data or the
       age of the sensor.

If 2 bytes of SN are sent using a counter, there is little to no leakage 
of sensor age.

If little traffic from sensor then only 1 byte may be better for this 
purpose.

I just don't see this as a risk if care is taken.  You may want to say this.

Finally where is the open source code available?

You need a UDP app in transport mode example in App 1.  :)

If you get this draft active, I will work on providing that example.  ;)


thanks.