Computing Vector Functions on Anonymous Networks

In an anonymous network processors have no distinct identifiers. For any class of anonymous networks, we characterize effectively the vector functions f : n → n which are computable in the class. Our results hold for both unidirectional and bidirectional networks, (partially) wireless networks, and three different processor activation models. We also identify the rigidity conditions under which a network can compute every function. Moreover, we improve the known bounds on the number of synchronous steps required to compute a function to n + , where n is the number of nodes of the network and its diameter, and show that this is optimal.

• Type: Conference
• Year: 1997
• Where: PODC
• Authors: Paolo Boldi, Sebastiano Vigna
• Keywords: Anonymous Network Processors, Anonymous Networks, Bidirectional Networks, Distributed And Parallel Computing, PODC 1997