Abstract
Issues regarding the effect of routing and load-balancing algorithms on the stability of a fault-tolerant distributed database system are examined. Reference is mainly to control algorithms. It is shown that the behavior of the routing algorithm is vital for the dependability of the system, and that the transitions induced by the algorithm should be stable or at least transitively stable. Degradation indexes are defined for the system dependability, and some open questions are raised.

This publication has 10 references indexed in Scilit: