From Key Predistribution to Key Redistribution
                    J.Cichon, Z.Golebiewski,M.Kutylowski
                           Wroclaw University of Technology


One of crucial disadvantages of key predistribution schemes for ad hoc 
networks is that if devices A and B usea shared key K to determine their
session keys, then any adversarial device that holds  K can impersonate 
A against B (or vice versa). Also, the adversary can eavesdrop communication 
between A and B for the lifetime of the system. 


We develop a dynamic scheme where  a system provider periodically 
broadcasts  random temporal keys (e.g. using GSM technology) encrypted 
with keys from the main  predistribution pool. Shared temporal keys (and not
the keys from the main pool) are used to establish session keys. 

The trick is that the scheme broadcast is organized in such a way that with 
a high probability two devices share much more temporal keys than the keys
from the main pool of keys. It is a kind of paradox, but this makes it possible
not only to protect communication against an adversary that has collected 
a large fraction of keys from the main pool, but also makes the system well 
suited for authentication purposes.  

ALGOSENSORS'2010, 
paper in LNCS Springer Verlag

work done within FRONTS Project