Hi Anass,

IMHO this can be handled by signing the WS-Addressing headers in the message.

For example in the given scenario, the message from Aclice to Bob will
have a "wsa:To" header which will hold the EPR of Bob. (Bob and
Charlie are two services hence they have differnt EPRs).
If Bob is to send a message to Charlie he will have in include
Charlie's EPR in the wsa:To header in his message to Charlie. And he
(Bob) will _not_ be able to do this and recompute the signature as
Alice.
Therefore the attack mentioned can be prevented.
Also to prevent replaying of the message Alice can add a wsu:Timestamp
and sign the wsu:Timestamp header as well.

Thanks,
Ruchith

On 3/2/06, anass merzak <[EMAIL PROTECTED]> wrote:
> hello all,
>
> I would like to know if wss4j is vulnerable to naive sign and encrypt trick,
> because according to
> http://world.std.com/~dtd/sign_encrypt/sign_encrypt7.html
> approximately, if Alice send a signed message to bob, can bob send the same
> message  to charlie  signed with alice  signature  (that it have receive),
> and thus make believe charlie that it is Alice which have send the message
>
> Also, would like to know youre response to
> http://neubia.com/archives/000363.html ,is wss4j immune
> against such tricks.
>
> thank you alot.
>
> --
>
> Anass Merzak
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to