Which of the following are insecure against replay attacks

[ecy5maa]

1. The problem statement, all variables and given/known data

 

Determine which of the following protocols are insecure

against replay attack , and explain why:

 

(a) A → B : username, password

 

(b) A → B : username, h(timestamp, password)

 

© A → B : username, timestamp, h(password)

 

(d) A → B : username, h(password), h(timestamp)

 

 

 

 

 

 

I know a is insecure, b is secure but I am not certain about c and d.

 

Off the cuff, in C it seems that h(password) can be replicated as its in the open, as can timestamp, but then if timestamp can be replicated so can h(timestamp). No?

 

I guess my answer is that a,c and d are all insecure, so if possible can someone provide any input on this?

[NetSplitter]

Interesting question! Here are my 2 cents

 

You are correct A. is vulnerable to replay attacks; it would take no effort atall.

 

B. Is more in the HMAC format of authentication, the hash is a timestamp +password. I would believe this to be secure. The attack would not be able toreverse the hash and change the time stamp, as this would be nearly impossible.

 

C. Insecure, there is no hash protecting the time stamp, so the timestamp couldbe changed theoretically to match the clock of the receiver.

 

D. Once again I think this would be secure, there would be no way to recomputethe hashed of the timestamp.

 

So for certain A and C is vulnerable to replay attacks.

 

I would consider B and D secure w/ the consideration that replay attacks arenot real time.

 

If interested read about Kerberos, it is a very time sensitive protocol usedwithin a MS AD network.

 

[ecy5maa]

But why would you need hashed value of the timestamp, is it not possible to just simply use h(timestamp) like we use h(password) in C?

 

 

Thanks for responding!

[NetSplitter]

Timestamps are time sensitive, passwords are not (unless changed of course, or one time key sessions).

 

I am not sure I quite understand what you are asking.

 

If the timestamp are not hashed the attacker could compute his own timestamp into the packet for a later time by comparing the time showed in the unhashed packet.

 

A hashed password doesn't protect against a replay attack because the attacker could replay the hash. The attack cannot do this with a timestamp because it will no longer be accept after its window is up.

[ecy5maa]

Timestamps are time sensitive, passwords are not (unless changed of course, or one time key sessions).

 

I am not sure I quite understand what you are asking.

 

If the timestamp are not hashed the attacker could compute his own timestamp into the packet for a later time by comparing the time showed in the unhashed packet.

 

A hashed password doesn't protect against a replay attack because the attacker could replay the hash. The attack cannot do this with a timestamp because it will no longer be accept after its window is up.

 

 

What I was asking was that if the non hashed time stamp can be used via manipulation, why not the hashed time stamp as well. But I read the book again and i have cleared up what a hashed value actually is. So i agree with you.....B&D are secure.

 

Thanks again for the help!