Hi,

>     > So in a way, a NOTIFY might make sense. But then we do run the risk of
>     > an IPsec SA being installed with the notify ignored using a default
>     > security context instead of the security context required. Having it
>     > as part of the TSi/TSr payloads makes that a lot more clear.
> 
>     > Thoughts?
> 
> It has to be a traffic selector.

I agree that it must be a traffic selector. However, in my opinion 
the way it is currently defined in the draft is wrong.

According to RFC7296 the individual TSs in TS payload are combined
using logical OR for the purpose of deciding whether given packet
matches the selector. In particular, Section 3.13 of RFC7296 says:

   ... a packet matches a given TSi/TSr if it matches at least one of the
   individual selectors in TSi, and at least one of the individual
   selectors in TSr.

The draft currently defines a new TS, that contains only security label
and no IP address, protocol and port information. Thus, it is impossible
with the current definition to express policy, that if packet is going
from IP1 to IP2 AND has a security label S, it must be protected 
using the negotiated SA. If you put IP1 and S in TSi and IP2 and S in TSr
it will mean that packet EITHER going from IP1 to IP2 (regardless
of its security label) or ANY packet having security label S (regardless 
of its IP addresses) is protected using the negotiated SA.

To fix this issue the draft should instead define two new traffic selectors - 
TS_IPV4_ADDR_RANGE_SECLABEL and TS_IPV6_ADDR_RANGE_SECLABEL,
that would contain IP range, protocol and ports range (as RFC7296 selectors),
 as well as the security label. This way the presence of the security label
will be an additional criteria for packet matching. 

Regards,
Valery.



_______________________________________________
IPsec mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ipsec

Reply via email to