2
$\begingroup$

I would like to know which method can provide better security: using identity based signature or hashing the identities and message (using SHA-1) for authentication?

$\endgroup$
2
  • $\begingroup$ Are you asking for a comparison of authenticating a message, comparing identity based signatures with something like SHA-1(SENDER_NAME || MESSAGE)? $\endgroup$
    – mikeazo
    Dec 14, 2015 at 19:33
  • $\begingroup$ yes, both SHA -1( sender identity || message) and using identity based signature can be used to authenticate the source and check data integrity, i would like to know which method is more secure? $\endgroup$
    – SARA
    Dec 15, 2015 at 6:12

2 Answers 2

1
$\begingroup$

Of the two, identity based is more secure. In fact, unless you have left details out of the question, a SHA-1 hash (or any hash for that matter) of the sender's identity and the message content provides no source authentication (or data integrity check).

According to your comment, the way you would use SHA-1 to authenticate this is to concatenate the senders identity and the message, and compute the SHA-1 hash of that value. The problem is, anyone who knows your identifier can do this for any message. To illustrate this, consider the message I, SARA, give all my money to mikeazo.. Surely you wouldn't send this message, but, as proof that you sent it, here is the "authentication" tag: 92a7904e083e41d8667b92b7100d9d970eb0e84c.

You can check that SHA-1(SARA I, SARA, give all my money to mikeazo.) == 92a7904e083e41d8667b92b7100d9d970eb0e84c. So that message must have come from you, so where is my money?

$\endgroup$
7
  • $\begingroup$ identity based os more secure just because the proposed hashing signature scheme provides no security at all :-) $\endgroup$ Dec 15, 2015 at 18:30
  • $\begingroup$ @mikeazo...In my project, I am encrypting the message and SHA-1( source ID, destination ID, message) by AES-128, so I can assume no adversary can know my message and thereby cannot generate hash, hence authenticating the source. Moreover in your example ,two messages give same hash, is it a case of hash collision? $\endgroup$
    – SARA
    Dec 16, 2015 at 9:48
  • $\begingroup$ @SARA, it is impossible for us to help you if you leave crucial details out of your question. Where does the key come from that you are encrypting with? Who else has the key? Who are you trying to prove authenticity to? $\endgroup$
    – mikeazo
    Dec 16, 2015 at 13:04
  • $\begingroup$ @mikeazo .apologies for incomplete information. consider 2 processors, one being the sender and the other receiver. The communication between them is automatic. both the sender and the reciever have same pre -installed 128 bit session key within the machine. The sender encrypts the message and SHA-1( source ID,destination ID, message) together by AES-128 using the pre installed session key. The rx decrypts the msg by the same session key,and calculates SHA-1 (S. ID, dest ID, msg) and compares if the recieved hash is same as the calculated hash. same values proves the authentication of source? $\endgroup$
    – SARA
    Dec 16, 2015 at 14:45
  • $\begingroup$ @SARA, so the proof is to the receiver, correct? Since there is no timestamp, couldn't anyone replay the ciphertext? So it really only authenticates that the sender, at some point in the past, sent the message to the receiver, since the encryption key never changes. $\endgroup$
    – mikeazo
    Dec 16, 2015 at 15:07
0
$\begingroup$

identity based signature suffers of single point failure but you dont need a public key infrastracture (i.e. you dont need certificates). conversely, classic signature schemes need a pki in order to work properly.

hashing itself doesnt provide any signature security. anybody can hash a message using someone identity.

so the short answer is: hashing alone provides zero security w.r.t. signing, everybody can hash something using someone else identity as the second input.

the key point of signing is that sign can be done only by the sender but the verify can be done by everybody. it's the dual of asymmetric encryption: everybody can encrypt a message for you but only you can decrypt it

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.