Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Key Exchange Using Passwords and Long Keys by dfhdhdhdhjr


									 Key Exchange Using
      and Long Keys
            Vladimir Kolesnikov
               Charles Rackoff

Comp. Sci. University of Toronto
Communication Setting


      Full Control   Insecure
Secure Communication from
Shared Random Key
            Trusted Party
               k 2R DK      • Simple
                            • Very efficient

              k2 2R DK
            Trusted Party
Key Exchange (KE)
A protocol between two parties
 Both output (the same) randomly chosen k 2 DK

 Adv does not know anything about k even if it sees
  all other exchanged keys
 Adv cannot mismatch players
     If Alice instance ``thinks’’ she exchanged a key with Bob,
      then at most one instance of “Bob talking to Alice” may
      have the same key
     Players must have secret credentials
Defining KE
   Large amount of prior work
   An intuitive notion, but hard to define

   We want our definition to:
       Be intuitive and easy to use
       Reject “bad” protocols (allow powerful adversaries)
       Accept “good” protocols (avoid unnecessary
Simulation Style KE Definition
       Real               Ideal


   8      9
                     • Powerful
                     • But complicated
Game Style KE Definition

                                     Plays the game:

                            • challenge a completed
                              honest player

                         • Present either a key
                           or a random string
                         Adversary guesses which
                         • Should not do too well

         • Seems to be almost as powerful
         • Self-contained
         • Simpler
Our Setting
• Asymmetric – Server (e.g. Bank) and Clients

 • Large secure storage            • Key on storage card
   of credentials                      • can be lost or stolen
                                   • Memorized password
                                       • low entropy
                                       • guessing attack possible
  • if card not stolen
       • have full security. Password guessing not possible
  • If card is stolen, still have password security
Some of Related Work
   Hybrid model (C has a pwd and pk of S)
       Halevi Krawczyk 99, Boyarsky 99
   Simulation- vs game-style KE
       Simulation-style KE
           Shoup 99, Boyko MacKenzie Patel 00
           Universally Composable (UC) Canetti Halevi Katz
            Lindell MacKenzie 05
       Game-style KE
           Bellare Pointcheval Rogaway 00
Denial of Access (DoA) Attack
   In Password-Authenticated KE, it is
    necessary to stop service if “too many”
    password failures P?
       Adv can deny access for good guys
   We can protect against such attacks
       Require that Adv cannot cause P?, unless he
        stole key card
       Don’t know of previous formalizations of DoA
           Complements Denial of Service notion
Our Protocol

Note: No Mutual Authentication
Password updates
   Usually handled externally to the definition
   If C updates his pwd, then DoA attack is
    possible (Adv can replay old msgs)
       Problem: have users with related credentials
   Solutions
       Update long key as well
       Have a challenge-response protocol
       Keep password update counters
       In the last two cases also need to update definition
Can a definition allow for
mistyping passwords?
   We don’t model this
   What if we allowed Adv to create instances
    with mistyped passwords?
       Adv specifies the password
           Is this how people mistype?
             can behave badly on pwd’ = pwd+1
       Adv specifies a mistyping function
           Only f that has 0,1,|D|-1 or |D| fixed points is allowed
   UC-based definitions can handle this
Definitional Choices: Counting
passwords attacks
   Adv can guess passwords
   Quantify advantage; “password attack”
   Previously
       Act of Adv interfering with traffic
       (Insignificant change? Successful guess?)
   In our definition
       Count failed password attacks – player outputs P?
   Define Key Exchange (KE) in a new model
       Generalization of the hybrid model of Halevi-
        Krawczyk (HK)
       (Some of) our discussion applies to other models
        (password-only and hybrid model of HK)
   Give a new efficient KE protocol
   Discuss a potential flaw in the HK protocols
       Some members of the family of the HK protocols
        are vulnerable to password guessing attacks
Extended version is on Eprint. Contains:
     Proofs
     Discussion on storing passwords on the server
     Discussion on password updates


To top