RE: I-D ACTION:draft-cruickshank-ipdvb-sec-00.txt, "ULEsec"

[Frank <frank-dlr-internet@online.de>]

> -----Original Message-----
> * The ipdvb security ID is focused on ULE message encryption and
> optional authentication.  
Yes, of course, the symmetric encryption and the key management
should be as isolated as possible. Something for future I-D :-)

> idea of having ULE header extensions for security, ULE MAC address
> hiding using unique security ID (like the SPI in IPsec) with
> or without
> temporary MAC address and the general requirements for ULE
> encryption .

Identity protection also was an issue in the UMTS security architecture
(IMSI vs. TMSI). This could be a requirement from governmental users.
However, take care that the identity (or equivalent permanent MAC/NPA
addresses) is protected during the whole communication session, including
the connection setup signaling. Encrypting the MAC after it was revealed
during setup does not help very much. Do the SNDU data formats
(Figure 1/2) allow such a procedure?

Your idea in the draft to link the temporary MAC to the key-exchange
protocol sounds reasonable. Otherwise you would need two independent
"re-keying" protocols for moving "targets", one for the dynamic MAC
and one for the symmetric encryption key.

MAC hiding using the SID sounds reasonable after setup. But how do you
want to prevent the (permanent)MAC disclosure during connection setup?

In general, I guess this draft is a good start for the IETF in Paris.
Have fun!

Frank