Article View: sci.crypt
Article #2749Re: Anti-spoofing protocol?
From: Carl Ellison
Date: Mon, 04 Jan 1993 23:30
Date: Mon, 04 Jan 1993 23:30
65 lines
3176 bytes
3176 bytes
In article <1993Jan4.215027.17258@netcom.com> rcain@netcom.com (Robert Cain) writes: >This all assumes that you have a secure means of getting your public key >to the other or an authenticated one. Both seem equivalent in the end >to swapping a secret key to me. This makes me question the fundamentals >of PK actually. PK doesn't allow absolute determination of identity. There is no such thing. What PK does (eg., via RSA) is put me in a private room which can not be eavesdropped with some one person. That one person is free to carry a wire and to lie about his name, age, even gender. That person is free to just mouth words supplied by someone else. That person is free to be a spoofer. Of course, this dishonest person is free to do all these things in the flesh. He doesn't need PK to do any of this. However, if that person is honest, then I'm having a private conversation. Through that conversation, I can get to know this person. What a shared secret can do (iff the person is honest) is establish a relationship between two events -- in this case, the opening of a private channel and the passing of a shared secret -- so that, if the person is honest, I'm assured that I'm speaking to the same person I spoke to in the earlier incident. The same can be done with PK -- under one RSA key, I contact someone, build a friendship and give him a secret to hold for future identification. Under a new RSA key, I encounter someone who claims to be my old friend. Within a private channel established to this possibly new person, I ask for evidence of the shared secret -- and when I get it, I know that these channels are probably to the same person. ------------------------------ An anti-spoofing protocol may be impossible to invent. It would have to tell the difference between: Case I: Alice calls Bob, sets up a secure channel by exchanging public keys over the line. Unknown to Alice, Eve is spoofing this. Alice <-> Eve <-> Bob Case II: Alice calls Bob, sets up a secure channel. Unknown to Alice, Bob has his assistant Carl handle all his channels to women with names A-M, so Bob reencrypts in Carl's key... Alice <-> Bob <-> Carl The first is spoofing by an eavesdropper. The second isn't. The block diagrams are the same. Can any protocol detect the difference, assuming that Alice and Bob have no shared secret (no knowledge of each other beforehand)? -- -- <<Disclaimer: All opinions expressed are my own, of course.>> -- Carl Ellison cme@sw.stratus.com -- Stratus Computer Inc. M3-2-BKW TEL: (508)460-2783 -- 55 Fairbanks Boulevard ; Marlborough MA 01752-1298 FAX: (508)624-7488
Technical Details
Message-ID:
<1iahbgINNt89@transfer.stratus.com>Path:
arc1-ca.pugleaf.net!sparky!uunet!noc.near.net!transfer.stratus.com!ellisun.sw.stratus.com!cmeReferences:
<1993Jan4.104051.23477@qualcomm.com> <1993Jan4.215027.17258@netcom.com>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!noc.near.net!transfer.stratus.com!ellisun.sw.stratus.com!cme From: cme@ellisun.sw.stratus.com (Carl Ellison) Newsgroups: sci.crypt Subject: Re: Anti-spoofing protocol? Date: 4 Jan 1993 23:30:56 GMT Organization: Stratus Computer, Software Engineering Lines: 65 Message-ID: <1iahbgINNt89@transfer.stratus.com> References: <1993Jan4.104051.23477@qualcomm.com> <1993Jan4.215027.17258@netcom.com> NNTP-Posting-Host: ellisun.sw.stratus.com