Mutual authentication using symmetric crypto: ISO 9798-2 versusMAP2
From: Ernest Hammingweight (hammingweight_at_hotmail.com)
Date: 12/29/04
- Next message: infobahn: "Re: code cracking or how do you know you've got the correct key?"
- Previous message: Phil Carmody: "Re: code cracking or how do you know you've got the correct key?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: 29 Dec 2004 02:17:22 -0800
Hi
I need a client and a server to authenticate each other. Both are
fairly limited devices (the client is a smartcard). I'd like to use a
symmetric crypto protocol. Two alternatives would appear to be the
MAP2 protocol of Bellare-Rogaway or the 3-pass protocol of ISO 9798-2.
I was wondering whether anybody had any suggestions which might be
better and any info on what authentication protocols are widely
deployed. Life is always easier if I can point to a precedent in
justifying my choice.
For those unfamiliar with the protocols they follow later (they're
fairly similar). E_K denotes some function using a key K and a
symmetric crypto algorithm that provides data authentication.
For both protocols the parties involved are A and B. R_A and R_B are
noces generated by A and B. Text1, Text2 and Text3 are arbitrary
strings that may be authenticated.
The MAP2 protocol is as follows:
A->B: R_A, Text1
B->A: E_K( B,A, R_A,R_B, Text1, Text2 )
A->B: E_K( A, R_B, Text3 )
ISO 9798-2 (slightly simplified) is as follows:
A->B: R_A, Text1
B->A: E_K( R_A, R_B, B, Text2 )
A->B: E_K( R_B, R_A, Text3 )
MAP2 looks superior to me. Why doesn't B authenticate the data Text1
in ISO9798-2? Also party B doesn't indicate that it knows it's
communicating with party A. In practice neither of these may be
serious (they're probably not for my purposes) but they suggest that
MAP2 is a better designed protocol. Another feature I like of MAP2 is
that it can easily be transformed into AKEP1 or AKEP2 which provide
authentication and key exchange while 9798-2 doesn't concern itself
with key exchange (and key exchange may be a 'nice-to-have').
Anyway, I'd appreciate all comments even if they just support or rebut
my prejudices.
Any indication that ISO 9798-2 is actually used would be appreciated.
- Next message: infobahn: "Re: code cracking or how do you know you've got the correct key?"
- Previous message: Phil Carmody: "Re: code cracking or how do you know you've got the correct key?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
|