Coordination Without Prior Agreement

Abstract

Assuming that there is an a priori agreement between processes on the names of shared memory locations, as done in almost all the publications on shared memory algorithms, is tantamount to assuming that agreement has already been solved at the lower-level. From a theoretical point of view, it is intriguing to figure out how coordination can be achieved without relying on such lower-level agreement. In order to better understand the new model, we have designed new algorithms without relying on such a priori lower-level agreement, and proved space lower bounds and impossibility results for several important problems, such as mutual exclusion, consensus, election and renaming. Using these results, we identify fundamental differences between the standard model where there is a lower-level agreement about the shared register's names and the strictly weaker model where there is no such agreement.