Otway Rees Protocol
Otway Rees Protocol
Verification
Li Chen 1,2
1. Computer Centre, Henan University of Finance and Economics,
Zhengzhou, Henan 450002, China
2. Department of Network Engineering, Information Engineering University,
Zhengzhou, Henan 450002, China
Abstract- Authentication protocol is applied to implement the but also is easily extended and has more sound semantics than
identity authentication of two communicating entities and build BAN-like logic.
secure communication tunnel through exchanging keys. The In the paper, we employ the SVO logic to verify the
paper analyzes the security properties of the original Otway security of the Otway Rees protocol and propose a new
Rees protocol and its existing improved version, it finds that the improved protocol.
protocol cannot meet the authentication goals. The paper Rest of the paper is organized as follows -in the next
proposes a novel improved protocol, which eliminates
section we present the Otway Rees protocol and analyze
redundancy in the Otway Rees protocol messages and removes
the limitations that the Otway Rees protocol cannot resist security of the protocol. In section 3, the improved Otway
impersonation attack by modifying message format and adding Rees protocol is proposed and verified by using the SVO
handshake message. The paper also verifies the security of the logic. Section 4 concludes the paper.
improved Otway Rees protocol by utilizing formal method. The
analysis results show that the protocol satisfies the strong goals
II. THE OTWAY REES PROTOCOL AND ITS ANALYSIS
of identity authentication and key distri bution. 2.1. Otway Rees Protocol
I. INTRODUCTION The Otway Rees protocol is a key distribution protocol
With the explosion of the Internet, electronic transactions and also guarantees authentication. It assumes a shared-key
have become more and more common. However the cryptosystem, in which each participant shares a master key
transactions' security is crucial to many applications, e.g. with a trusted party, the Key Distribution Sever. The Otway
electronic commerce, digital contract signing, electronic Rees protocol is described as follows.
voting, and so on. While issues such as confidentiality, (1) A-7B : M, A, B, [NA , M, A, B]KAS
authentication, access control, etc. have been studied (2) B-7S : M, A, B, [NA , M, A, B]KAS, [NB , M, A, B]KBS
intensively, most interest in authentication protocol has only
come in recent years. (3) S-7B : M, [NA , KAB]KAS, [NB , KAB] KBS
Authentication protocols are not only used to implement (4) B-7A : AI, [NA , KAB]KAS
the identity authentication of two communicating entities, but
also used to build secure tunnel through exchanging keys
between two communicating entities. In fact, many
authentication protocols, such as Otway Rees protocol [1], (3)M [NA. KAS]KAS'
Kerberos protocol [2] and Internet Key Exchange protocol [NB. KBs
(2) M. A. B, [NA. M. A,
(IKE) [3], are all designed to both of the above two [NB. M A. B]KBs
requirements. However it is very difficult to design the
protocols meeting the above requirements because the
potential security vulnerabilities are covert. For example, f;\ (I)M.A,H,{NA ,M.A.B]KAS
(4)M [NA.
Reference [4-5] found that the Otway Rees protocol could not
resist impersonation attack, etc. Fig. 1. The Otway Rees protocol
Formal methods have been widely used to verify the
security of authentication protocols in recent years. Many In the Otway Rees protocol, initially, the protocol initiator
significant results have been achieved in the area since formal A and the other participant B share keys KAS and KBS with the
methods began to apply to cryptographic protocol security server S respectively. A firstly sends to B message M, her own
analysis. The SVO logic [6,7] used for authentication protocol identifier A, B's identifier B and an encrypted chunk [NA , M,
analysis is a many-sorted modal logic, which captures the A, B]KAS , where NA is A's nonce. In the second message, Bob
desirable properties of BAN-Like logic, such as BAN logic sends to the server S message M, A's identifier A, his own
[8], GNY logic [9], AT logic [10] and VO logic [11]. It has identifier B and two encrypted chunks [NA , M, A, B]KAS , [NB ,
not only better linguistic expressibility and logical derivability,
499
Authorized licensed use limited to: Manchester Metropolitan University. Downloaded on November 30,2020 at 15:01:23 UTC from IEEE Xplore. Restrictions apply.
AI6 <t> PI4 B believes B received (M, A, <[NA, M, BjKAS >*B)
(7) Freshness axiom PI5 S believes S received (M, A, B, <[NA, M, BjKAS,
AI7 fresh fresh (Xl, ... , X n ) [NB, M, AjKBS>*s)
(8) Nonce-verification axiom PI6 B believes B received (M, <[NA, B, KAB]KAS>*B,
A 19 (fresh ( X) /\ P said X ) P says X [NB, A, <KAB>*BjKBS )
3) SVO Logic Definitions P 17 A believes A received (M, [NA, B, <KAB >*A]KAS,
P Q : K is a good key for P and Q regardless of [NB, Bj <KAB>*A)
whether either of them knows it. PI8 B believes B received (M, [NB-l, Aj <KAB>*B)
In addition, a common consequence can be inferred from 5) IORP Interpretation Assumptions
axiom A 1 and rule MP.
PI9 B believes B received (M, <[NA, B, KABjKAS>*B,
Al +MP P believes <t> /\ P believes (<t> tJ!) (P
believes \}l) [NB, A, <KAB>*BjKBS ) B believes B received
(M, <[NA, B, KABjKAS>*B, [NB, A, A
Because AO cannot be deduced from other axioms and it
( <KAB>*B ) BjK )
must be used in the analysis process later. We also need to BS
add the believing axiom AO to the SVO logic. P20 A believes A received (M, [NA, B, <KAB>*AjKAS,
AO (P believes <t> /\ P believes \}l) equivalent (P [NB, Bj <KAB>*A) A believes A received (M,
believes <t> /\ \}l) [NA, B, A ( <K..w>*A ) B,fresh«KAB>*AjKAs, [NB,
3. 3. Security Verification ofthe IORP Protocol Bj <KAB>*A)
Now we verify the security of the improved Otway Rees P2I A believes A received ([NB, B] <KAB>*A) /\ (A
protocol by utilizing the SVO logic. In every step of the believes A ( <K..w>*A ) B) A believes A
analysis process, we give the inference result, and then give
the inferences rules, axioms, definitions, assumptions and received ([<NB>*A, B, A ( <K.W>*A ) B] <KAB>*A)
formulae, which are required when we infer the result. P22 B believes B rece ived ([NB-l,A] <KAB >*B) /\ (B
1) IORP Goals believes A ( <KAS >*8 ) B ) B believes B
GI A believes A ( K A8 ) B received ([NB-l, A, A ( <K..W>*B ) B ]<KAB>*B)
G2 A believes fresh(KAB ) 6) IORP Derivation for A
G3 A believes B believes A ( K ..w ) B (1) A believes A received (M, [NA, B, A ( <K..w>*4 ) B,
G4 A believes B believes fresh(KAB ) fresh«KAB>*AjKAs, [NB, Bj <KAB>*A)
G5 B believes A ( KjR ) B By AO, MP, PI7, P20
G6 B believes fresh(KAB ) (2) A believes A received (M, [NA, B, A ( <KAB>*A ) B,
G7 B believes A believes A ( K AB ) B fresh (<KAB>*AjKAS )
G8 B believes A believes fresh(KAB ) By AI+MP, A7, (1)
2) IORP Initial Assumptions (3) A believes S said (M, [NA, B, A ( <K.W>*A ) B,
fresh (<KAB>*A])
PI A believes A ( K AS ) S
By AO, A3, PI, (2)
P2 B believes B ( K BS ) S (4) A believes Ssays (M, [NA, B, A ( <K..w>*A ) B,
P3 A believes S controls A B fresh (<KAB>*Aj)
By AO, A 17, AI9, P5, P7, (3)
P4 B believes S controls A B
(5) A believes A ( <KAB>*A ) B
P5 A believes S controls fresh(A ( K ..w ) B)
By AO, AI4, AI6, P3, (4)
P6 B believes S controls fresh(A ( K ..w ) B) (6) A believes fresh (<KAB>*A)
P7 A believes fresh(NA) By AO, AI7, AI9, P5, P7, (4)
P8 B believes fresh(NB ) (7) A believes A received ([NB, B] <KAB>*A)
By AI+MP, A7, (1)
3) IORP Received Message Assumptions
(8) A believes A received ([<NB>*A, B, A ( <K.w >*1 ) B]
P9 B received (M, A, [NA, B, M]KAs )
<KAB>*A)
PIO S received (M, A, B, [NA, M, B]KAS, [NB, M, A]KBS ) By Al +MP, P2I, (5), (7)
PI 1 B received (M, [NA, B, KAB]KAS, [NB, A, KAB]KBS ) (9) A believes B said ([<NB>*A, B, A ( <KAB>*A ) Bj)
PI2 A received (M, [NA, B, KABjKAS, [NB, BjKAB ) By AO, A3, P21, (5), (8)
PI3 B received (M, [NB-l, AjKAB ) (IO)A believes B has <KAB>*A
4) IORP Comprehension Assumptions By AO, Al+MP, A3, P21, (5)
(Il)A believes B says ([<NB>*A, B, A ( <KdR>*A ) Bj)
500
Authorized licensed use limited to: Manchester Metropolitan University. Downloaded on November 30,2020 at 15:01:23 UTC from IEEE Xplore. Restrictions apply.
By AO, AI7, AI9, (9), (10) protocol meets the strong authentication goals.
(12) A believes B believes A ( <KAB>*A ) B IV. CONCLUSION
By AO, AI4, AI6, (11)
With the phenomenal growth of the Internet and open
(13) A believes B believesfresh(<KAB>*A) networks in general, security services, such as identifY
By AO, AI7, AI9, P7, (11) authentication and key distribution, become crucial to many
On the basis of the above analyses, we can conclude that applications. Authentication protocols are the important
the Otway Rees protocol satisfies the authentication goals for security protocols, and their security must be analyzed strictly
A. by using the formal methods. The SVO logic is an efficient
7) IORP Derivation for B formal method for analyzing authentication protocols. The
paper analyzes the security of the Otway Rees protocol. In
(1) B believes B received (M, <[NA, B, K AB]KAS>*B, [NB, addition, an improved protocol is proposed and its objective
A, A ( <K.W>*B ) B]KBs ) and security are formally verified.
By AO, MP, PI6, PI9
REFERENCES
(2) B believes B received ([NB, A, A ( <K,.W>*B ) B]KB:J
[1]D OTWAY, 0 REES. "Efficient and timely mutual authentication".
By AO, AI+MP, A7, (1) Operating Systems Review, 1987,21 (1): 8 - 10.
(3) B believes S said (NB, A, A ( <K..W>*B ) B ) [2] JT Kohl, BC Neuman. "The Kerberos Network Authentication Service".
RFC1510, 1993.
By AO, A3, P2, (2) [3] D Harkins, D Carrel. "The Internet Key Exchange (IKE)". RFC2.J09, 1998.
(4) B believes S says (NB, A, A ( <K.W>*B ) B) [4] C. Boyd, W. Mao, "Limitations of Logical Analysis of Cryptographic
Protocols", EUROCRYPr93.
By AO, AI7, AI9, P6, P8, (3) [5] M. Abadi, R. Needham. "Pmdent engineering practice for cryptographic
(5) B believes A ( <KAB>*B ) B protocols". Proceedings of 1994 IEEE Symposium on Research in Security
and Privacy, Oakland, California, May 1994, pp 122-136.
By AO, AI4, AI6, P4, (4) [6] F Paul. Syverson, C Paul. van Oorschot. "On Unifying Some
(6) B believes fresh( <KAB>*B ) Cryptographic Protocol Logics". Proceedings of the IEEE Computer
Society Symposium in Security and Privacy in Los Alamitos, 1994,
By AO, AI7, AI9, P6, and (4) pp.14-28.
(7) B believes B received ([NB-I,A] <KAB>*B) [7] 1. Wen, M. Zhang, X. Li. "The study on the application of SVO logic in
By AO, AI+MP, A7, PI8 formal analysis of authentication protocols". Proceedings of the 7th
(8) B believes B received ([NB-I, A, A international conference on J'}ectronic commerce.2005, Vol. 113,
pp.744-747
( <KAB>*B ) B ]<KAB>*B) [8] M. Burrows, M.Abadi, R. Needham. "A logic of authentication". ACM
By AI+MP, p22, (5), (7) Transactions on Computer Systems (lDeS), 1990, vol.8, no. 1, pp.I8-36.
[9] L Gong, R Needham, R Otway Rees. "Reasoning about belief in
(9) B believes A said (Ns-I, A, A ( ) B) cryptographic protocols". In: Proceedings of the IEEE Computer Society
By AO, A3, (5), (8) Symposium on Research in Security and Privacy. Los Alamitos: IEEl,-'
Computer Society Press, 1990, pp.234-248.
(10) B believes A says (NB-I, A, A ( <K.W>*B ) B) [10] M Abadi, MR Tuttle. "A semantics for a logic of authentication". In:
By AO, AI7, AI9, (8), (9) Proceedings of the 10th ACM Symposium on Principles of Distributed
Computing. ACM Press, 1991, pp. 201-216.
(11) BbelievesA believes A ( <KAB>*H )B [11] PC Van Oorschot. "Extending cryptographic logics of belief to key
By AO, AI4, AI6, (10) agreement protocols". In: Proceedings of the 1st ACM Conference on
Computer and Communications Security. ACM Press, 1993, pp. 233-243.
(12) B believes A believesfresh(<KAB>*B)
By AO, AI7, AI9, P8, (10)
The above analyses show that the improved Otway Rees
501
Authorized licensed use limited to: Manchester Metropolitan University. Downloaded on November 30,2020 at 15:01:23 UTC from IEEE Xplore. Restrictions apply.