hhow09's Blog

Protecting Replay Attack

Replay Attack #

An attack in which the Attacker is able to replay previously captured messages (between a legitimate Claimant and a Verifier) to masquerade as that Claimant to the Verifier or vice versa.

Protection Approaches #

  1. IP rate limiting (extra cost)
  2. Enterprise solution (extra cost)
  3. API key (not safe for public client)
  4. nonce (with authentication)

Simple solution without extra cost: Cryptographic nonce #

Nonce in cryptography means “number once,” and this arbitrary number is only used one time in a cryptographic communication.

The nonce helps to prove that the message received was sent by the intended sender and was not intercepted and resent by a bad actor.

Nonce Auth

Typical client-server communication during a nonce-based authentication process including both a server nonce and a client nonce.

How to choose a nonce #

  1. Timestamp

    • client use timestamp as nonce in the request
    • server should verify the timestamp within a certain range
    • beware of client time skew
      • user could change the device time, therefore we could use timestamp returned from server.

  2. random number

    • client generates a random number as nonce
    • server checks the nonce is not used before
    • cons: need to store nonce in server for some time (e.g. in cache)

Is nonce enough? #

using nonce without encryption or authentication is easy to be guessed by attacker.

Encryption / Authentication #

Solution I used #

  1. client has public key from server
  2. client receive timestamp from server as nonce
  3. client encrypt the nonce with public key and send to server
  4. server decrypt the nonce with private key and verify the timestamp within a certain range.

Reference #