Master of Science (MS)
First Committee Member
Ajoy K. Datta
Number of Pages
We propose an algorithm for simulating atomic registers, test-and-set, fetch-and-add, and read-modify-write registers in a message passing system. The algorithm is fault tolerant and works correctly in presence of up to (N/2) -1 node failures where N is the number of processors in the system. The high resilience of the algorithm is obtained by using randomized consensus algorithms and a robust communication primitive. The use of this primitive allows a processor to exchange local information with a majority of processors in a consistent way, and therefore to take decisions safely. The simulator makes it possible to translate algorithms for the shared memory model to that for the message passing model. With some minor modifications the algorithm can be used to robustly simulate shared queues, shared stacks, etc. (Abstract shortened with permission of author.).
Distributed; Memory; Sharing; System
University of Nevada, Las Vegas
If you are the rightful copyright holder of this dissertation or thesis and wish to have the full text removed from Digital Scholarship@UNLV, please submit a request to firstname.lastname@example.org and include clear identification of the work, preferably with URL.
Aguilar, Oscar Rodrigo, "Sharing memory in distributed systems" (1990). UNLV Retrospective Theses & Dissertations. 92.