Recent literature presents several rooted tree based member deletion/revocation schemes trying to simultaneously minimize the key storage whileproviding efficient member deletion/revocation. Many of these approaches have different solutions and provide different values for the number of keys to be stored and distributed.

In this paper, we show that these problems can be systematically studied using basic concepts from information theory. In particular, we show that the entropy of member revocation event plays a major role in defining the key allocation requirements. We then relate the entropy of member revocation event to bounds on the key length.

We also show that the optimal Huffman coding strategy used in leads to security weaknesses. A method for generating key management schemes to withstand varying degrees of member collusion is also presented.

ATIRP 4th Annual Conference