This page should provide some additional information on security protocols and how to analyse and attack them.
Note: This page is still work in progress. Any comments, remarks, suggestions or requests are welcome and can be send to Pim Vullers.
A security protocol is normally given in the following notation:
A -> B : m. Here A stands for Alice, in this case
the sending agent; B stands for Bob, in this case the receiving
agent; and m is the message sent from Alice to Bob.
Other agents that might occur are: C for Charlie in three-party
protocols or Client in client-server protocols; S for Server in
client-server protocols; and T for Trudy or Trusted third party.
E for Eve, Eavesdropper or Evil attacker; and sometimes
M for Mallory or the Man-in-the-middle.
Eavesdropping is a passive attack (on confidentiality) in which Eve only listens in on a conversation between Alice and Bob. This can be countered by applying encryption on the messages sent between Alice and Bob.
References:
In a replay attack a previously eavesdropped message from Alice to Bob is sent again to Bob. For example logging in to an account using a username and password acquired by eavesdropping on a previous login session. This can be countered by properly using a nonce or counter within the encryption to ensure freshness of the messages.
References:
A reflection attack is used to attack challenge-response authentication protocols. When Eve receives a challenge from Bob, she will start a new session in which she sends this challenge back to Bob. The response she gets from Bob can now be used in the original session to convince Bob of her authenticity, while in fact Bob constructed the response himself. This can be countered by including the identity of the challenged agent into the challenge, this way Bob will recognise that the reflected challenge was not meant for him.
References:
A MITM attack is often used to attack key exchanges and public-key protocols. By sitting in between the two parties Mallory can modify the keys which are exchanged between Alice and Bob. Any further actions of Alice and Bob using these keys are compromised, which makes this a very powerful attack. Wikipedia provides a good example on how this attack can be used.
A passive MITM attack in which Mallory does not modify the messages is called relaying and is, in general, not considered as an attack. See below for more information.
References:
This is a special case of MITM attacks in which the man in the middle only forwards the messages. Relaying is in general not considered as an attack since this is what happens on networks all the time, messages are forwarded from one place to the other.
However, in special cases, relaying can create some attack scenarios. Especially when the communicating parties are assumed to be close to each other, for example a remote car key or an access card to enter the building. In these systems relaying of messages is an undesired action since it breaks the assumption that the owner of the token (the key or card) is close to the asset (the car or building).
References:
A denial-of-service attack (DoS attack) or distributed denial-of-service attack (DDoS attack) is an attempt to make a computer resource unavailable to its intended users, and thus attacks the availability property. This kind of attack does not have a specific method but usualy involves overloading the resource or service (with data or requests). This makes it also difficult to protect against this kind of attacks.
References: