review

From: Ioannis Giotis (giotis@cs.washington.edu)
Date: Tue Nov 23 2004 - 16:57:51 PST

  • Next message: Danny Wyatt: "Intercepting Mobile Communications: The Insecurity of 802.11"

    Security in wireless systems is a more challenging problem than in wired
    networks as adversaries can easily listen all traffic. 802.11 uses WEP
    as a security protocol. WEP is a secret key cryptosystem that uses the
    secret key and a random vector to generate an encoding key of equal
    length to the packet.

     

    Unfortunately, despite the desired theoretical aspects of WEP, poor
    implementations and practice, render this system somewhat insecure.
    First of all, network card manufacturers did not use much randomness in
    creating the "random" vector making it easy to find two packets encoded
    using the same vector. Secondly, using known dictionary techniques one
    could take advantage of partial knowledge of the packet's contents to
    gain full access to all packets. Finally, as in all secret key systems,
    the distribution of the secret key itself renders this system somewhat
    insecure.

     

    Another aspect of WEP is message authentication and digital signatures.
    Using a CRC-32 checksum, encoded with the message, is not enough to
    guarantee malicious attacks detection. Since CRC-32 was built as a
    random error detection mechanism, it can be exploited by someone who
    already knows its nature and is careful enough.

     

    I was surprised to find the same old mistakes done in older secret key
    cryptosystems, found in WEP as well. The authors do a good job at
    pointing out these problems, and their claims seam reasonable. However,
    the paper leaves something to be desired. By 2001, there were a lot more
    common problems when using secret keys, that I don't see the authors
    mentioning. One could also try to exploit several other properties of a
    wireless network. For example, could one exploit network protocol
    mechanisms to derive the secret key by say jamming frequencies ?

     

    Nothing much has changed since the paper was written. The same problems
    remain and we are clearly away from finding robust solutions in these
    area. I would imagine that large deployment of wireless networks these
    days will bring surface more and more ways to attack WEP.




  • Next message: Danny Wyatt: "Intercepting Mobile Communications: The Insecurity of 802.11"

    This archive was generated by hypermail 2.1.6 : Tue Nov 23 2004 - 16:57:52 PST