Secure distributed computing addresses the problem of per- forming a computation with a number of mutually distrustful participants,
in such a way that each of the participants has only limited access to the information needed for doing the computation. In
the presence of a third party, completely trusted by all participants the problem is trivially solvable. However, this assumption
is in many applications non-realistic. Over the past two decades, a number of solutions requiring no trusted third party have
been developed using cryptographic techniques. The disadvantage of these cryptographic solutions is the excessive communication
overhead they incur.
In this paper, we will show how to overcome these disadvantages and thus enable secure distributed computations in practice.
Our approach uses mobile agents employing these cryptographic techniques to provide for a trade-off between communication
overhead and trust. The communication overhead problem would be solved if the communicating parties were brought close enough
together. Our solution is to use mobile agents to execute the cryptographic protocols. Of course, a mobile agent needs to
trust his execution platform, but we show that the trust requirements in this case are much lower than for a classical trusted
third party.
A preliminary version of this paper was presented at the ECOOP 2000 Workshop on Mobile Object Systems
Postdoctoral Fellow of the Belgian National Fund for Scientific Research (F.W.O.)