Summary: Wait­Free Algorithms for Fast, Long­Lived Renaming \Lambda
Mark Moir and James H. Anderson
Department of Computer Science
The University of North Carolina at Chapel Hill
Chapel Hill, North Carolina 27599­3175, USA
August 1994
Abstract
We consider wait­free solutions to the renaming problem for shared­memory multiprocessing systems
[3, 5]. In the renaming problem, processes are required to choose new names in order to reduce the
size of their name space. Previous solutions to the renaming problem have time complexity that is
dependent on the size of the original name space, and allow processes to acquire names only once. In
this paper, we present several new renaming algorithms. Most of our algorithms have time complexity
that is independent of the size of the original name space, and some of our algorithms solve a new, more
general version of the renaming problem called long­lived renaming. In long­lived renaming algorithms,
processes may repeatedly acquire and release names.
1 Introduction
In the M ­renaming problem [2], each of k processes is required to choose a distinct value, called a name,
that ranges over f0; :::; M \Gamma 1g. Each process is assumed to have a unique process identifier ranging over
f0::N \Gamma 1g. It is further required that k Ÿ M ! N . Thus, an M­renaming algorithm is invoked by k
processes in order to reduce the size of their name space from N to M .

Source: Anderson, James - Department of Computer Science, University of North Carolina at Chapel Hill

Collections: Computer Technologies and Information Sciences