CONCLUSIONS

We have reported some recent results on the inference problem for additive queries on a statistical database. We met with some computationally hard problems (on strong safety and strong p-safety), which are likely to have no efficient solutions. If so, one should be content with weaker protection levels (such as weak safety and weak p-safety). In any case, we can conclude that much work still remains to be done for the security of statistical databases to be solved in a satisfactory way.