Noise Explorer beta

NX

a e

Message Pattern Analysis

Message A is the first message in the NX Noise Handshake Pattern. It is sent from the initiator to the responder. In this detailed analysis, we attempt to give you some insight into the protocol logic underlying this message. The insight given here does not fully extend down to fully illustrate the exact state transformations conducted by the formal model, but it does describe them at least informally in order to help illustrate how Message A affects the protocol.

Sending Message A

In the applied pi calculus, the initiator's process prepares Message A using the following function:

letfun writeMessage_a(me:principal, them:principal, hs:handshakestate, payload:bitstring, sid:sessionid) = let (ss:symmetricstate, s:keypair, e:keypair, rs:key, re:key, psk:key, initiator:bool) = handshakestateunpack(hs) in let (ne:bitstring, ns:bitstring, ciphertext:bitstring) = (empty, empty, empty) in let e = generate_keypair(key_e(me, them, sid)) in let ne = key2bit(getpublickey(e)) in let ss = mixHash(ss, ne) in (* No PSK, so skipping mixKey *) let (ss:symmetricstate, ciphertext:bitstring) = encryptAndHash(ss, payload) in let hs = handshakestatepack(ss, s, e, rs, re, psk, initiator) in let message_buffer = concat3(ne, ns, ciphertext) in (hs, message_buffer).

How each token is processed by the initiator:

If a static public key was communicated as part of this message, it would have been encrypted as ciphertext1. However, since the initiator does not communicate a static public key here, that value is left empty.

Message A's payload, which is modeled as the output of the function msg_a(initiatorIdentity, responderIdentity, sessionId), is encrypted as ciphertext2. This invokes the following operations:

Receiving Message A

In the applied pi calculus, the initiator's process prepares Message A using the following function:

letfun readMessage_a(me:principal, them:principal, hs:handshakestate, message:bitstring, sid:sessionid) = let (ss:symmetricstate, s:keypair, e:keypair, rs:key, re:key, psk:key, initiator:bool) = handshakestateunpack(hs) in let (ne:bitstring, ns:bitstring, ciphertext:bitstring) = deconcat3(message) in let valid1 = true in let re = bit2key(ne) in let ss = mixHash(ss, key2bit(re)) in (* No PSK, so skipping mixKey *) let (ss:symmetricstate, plaintext:bitstring, valid2:bool) = decryptAndHash(ss, ciphertext) in if ((valid1 && valid2)) then ( let hs = handshakestatepack(ss, s, e, rs, re, psk, initiator) in (hs, plaintext, true) ).

How each token is processed by the responder:

If a static public key was communicated as part of this message, it would have been encrypted as ciphertext1. However, since the initiator does not communicate a static public key here, that value is left empty.

Message A's payload, which is modeled as the output of the function msg_a(initiatorIdentity, responderIdentity, sessionId), is encrypted as ciphertext2. This invokes the following operations:

Queries and Results

Message A is tested against four authentication queries and five confidentiality queries.

Authentication Grade 1: Failed


RESULT event(RecvMsg(bob,alice,stagepack_a(sid_b),m)) ==> event(SendMsg(alice,c_1108,stagepack_a(sid_a),m)) || event(LeakS(phase0,bob)) cannot be proved.

In this query, we test for sender authentication and message integrity. If Bob receives a valid message from Alice, then Alice must have sent that message to someone, or Alice had their long term keys, which are not used, compromised before the session began, or Bob had their static key compromised before the session began.

Authentication Grade 2: Failed


RESULT event(RecvMsg(bob,alice,stagepack_a(sid_b),m)) ==> event(SendMsg(alice,c_1108,stagepack_a(sid_a),m)) cannot be proved.

In this query, we test for sender authentication and is Key Compromise Impersonation resistance. If Bob receives a valid message from Alice, then Alice must have sent that message to someone, or Alice had their static key compromised before the session began.

Authentication Grade 3: Failed


RESULT event(RecvMsg(bob,alice,stagepack_a(sid_b),m)) ==> event(SendMsg(alice,bob,stagepack_a(sid_a),m)) || event(LeakS(phase0,bob)) cannot be proved.

In this query, we test for sender and receiver authentication and message integrity. If Bob receives a valid message from Alice, then Alice must have sent that message to Bob specifically, or Alice had their long term keys, which are not used, compromised before the session began, or Bob had their static key compromised before the session began.

Authentication Grade 4: Failed


RESULT event(RecvMsg(bob,alice,stagepack_a(sid_b),m)) ==> event(SendMsg(alice,bob,stagepack_a(sid_a),m)) cannot be proved.

In this query, we test for sender and receiver authentication and is Key Compromise Impersonation resistance. If Bob receives a valid message from Alice, then Alice must have sent that message to Bob specifically, or Alice had their long term keys, which are not used, compromised before the session began.

Confidentiality Grade 1: Failed


RESULT attacker_p1(msg_a(alice,bob,sid_a)) ==> event(LeakS(px,bob)) cannot be proved.

In this query, we test for message secrecy by checking if a passive attacker is able to retrieve the payload plaintext only by compromising Bob's static key either before or after the protocol session.

Confidentiality Grade 2: Failed


RESULT attacker_p1(msg_a(alice,bob,sid_a)) ==> event(LeakS(px,bob)) cannot be proved.

In this query, we test for message secrecy by checking if an active attacker is able to retrieve the payload plaintext only by compromising Bob's static key either before or after the protocol session.

Confidentiality Grade 3: Failed


RESULT attacker_p1(msg_a(alice,bob,sid_a)) ==> event(LeakS(phase0,bob)) || (event(LeakS(px,bob)) && event(LeakS(pz,alice))) cannot be proved.

In this query, we test for forward secrecy by checking if a passive attacker is able to retrieve the payload plaintext only by compromising Bob's static key before the protocol session, or after the protocol session along with Alice's long term keys, which are not used, at any time.

Confidentiality Grade 4: Failed


RESULT attacker_p1(msg_a(alice,bob,sid_a)) ==> event(LeakS(phase0,bob)) || (event(LeakS(px,bob)) && event(LeakS(pz,alice))) cannot be proved.

In this query, we test for weak forward secrecy by checking if an active attacker is able to retrieve the payload plaintext only by compromising Bob's static key before the protocol session, or after the protocol session along with Alice's long term keys, which are not used, at any time.

Confidentiality Grade 5: Failed


RESULT attacker_p1(msg_a(alice,bob,sid_a)) ==> event(LeakS(phase0,bob)) cannot be proved.

In this query, we test for strong forward secrecy by checking if an active attacker is able to retrieve the payload plaintext only by compromising Bob's static key before the protocol session.