Hi, Nancy,
"Nancy Cam-Winget (ncamwing)" <ncamw...@cisco.com> 写于 2012-07-15 11:34:55: > > "Finally, it describes how the tls-unique channel binding [RFC5929] > may be used to PA-TNC exchanges > to the EAP tunnel method, defeating MITM attacks such as the > Asokan attack [Asokan]." > ==> The sentence is excerpt from the draft-ietf-nea-pt-eap-02, I wrote it down but forgot to write the comment. I was a bit confused by the solution to Asokan attack, because I found two solutions in two different places in the draft : 1. tls-unique 2.crypto binding > > "Some EAP tunnel methods may provide explicit confirmation of inner > method success; others may not. " > [NCW] I am not sure I understand the comment or request. The above > sentence is true; in > PT-EAP's case, as it is not an authenticating method, we describe > how tls-unique is used > in PT-EAP to address such binding. So, the sentence stands on its > own as it is reference > further details to follow. > This sentence is also excerpt from the draft-ietf-nea-pt-eap-02, again I forgot to write the comment :) . I don't see how current tunnel method provide explicit confirmation of inner method success. If it refers to the mixture of tunnel method key and inner method key, yes, there are some tunnel methods do this mixture (to against MITM attack ) and some do not do, but those tunnel method that do not mix the keys are not recommended to use Ｉ believe. > > "These inner methods may perform additional security handshakes including more > granular authentications or exchanges of integrity information (such > as PT-EAP.) " > ===> IMO，PT-EAP better be exchanged after the phase two of the EAP > tunnel method, so that > the resulted key derived from tunnel and inner authentication > method can be used to protect it. > [NCW] Do you mean to enforce an authentication (inner) method prior to PT-EAP? Yes. It seems to me that PT-EAP is an inner method provinding plain message transportation, not authentication. And EAP tunnel method is seldom used alone, becauce its purpose is to protect the inner authentication method, and only server authentication is provided in implemementation. so, I doubt the security of PT-EAP message transimitted in the EAP Tunnel method without another inner authentication method. Maybe in this case, mutual authentication could be enforced in EAP tunnel method, so that no inner authentication methods are needed. -Sujing Zhou
_______________________________________________ Emu mailing list Emu@ietf.org https://www.ietf.org/mailman/listinfo/emu
