1993•Systems and Computers in JapanRequires access

An improved √N algorithm for mutual exclusion in decentralized systems

Takeshi Fuchi

Open publisher page 0 citations

Abstract

Abstract An algorithm is proposed which realizes mutual exclusion in a computer network system where communications are conducted by messages only without using a shared memory. With this method, and by transferring a token representing the exclusion the exclusion right, the mutual exclusion is realized. A node wanting to execute a critical section first sends messages requesting exclusion to a set of nodes (members) assigned to it. The members keep this message, after receiving a finish exclusion message from some other node, transfer the exclusion request messages to the sender. The node receiving the token processes the critical section, and when finished, sends finish exclusion messages to the members. The members return information on nodes requesting exclusion known to them. According to this information, the next receiver of the token is determined. By contriving the selection of members and gathering information on nodes requesting exclusion, the mutual exclusion can be realized using √N + 1 or more but less than or equal to 3√N + 1 messages.

About this research paper

What this paper is about

Abstract An algorithm is proposed which realizes mutual exclusion in a computer network system where communications are conducted by messages only without using a shared memory. With this method, and by transferring a token representing the exclusion the exclusion right, the mutual exclusion is realized. A node wanting to execute a critical section first sends messages requesting exclusion to a set of nodes (members) assigned to it. The members keep this message, after receiving a finish exclusion message from some other node, transfer the exclusion request messages to the sender. The node receiving the token processes the critical section, and when finished, sends finish exclusion messages to the members. The members return information on nodes requesting exclusion known to them. According to this information, the next receiver of the token is determined. By contriving the selection of members and gathering information on nodes requesting exclusion, the mutual exclusion can be realized using √N + 1 or more but less than or equal to 3√N + 1 messages.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Abstract An algorithm is proposed which realizes mutual exclusion in a computer network system where communications are conducted by messages only without using a shared memory. With this method, and by transferring a token representing the exclusion the exclusion right, the mutual exclusion is realized. A node wanting to execute a critical section first sends messages requesting exclusion to a set of nodes (members) assigned to it. The members keep this message, after receiving a finish exclusion message from some other node, transfer the exclusion request messages to the sender. The node receiving the token processes the critical section, and when finished, sends finish exclusion messages to the members. The members return information on nodes requesting exclusion known to them. According to this information, the next receiver of the token is determined. By contriving the selection of members and gathering information on nodes requesting exclusion, the mutual exclusion can be realized using √N + 1 or more but less than or equal to 3√N + 1 messages.

Key concepts: Mutual exclusion, Suzuki-Kasami algorithm, Critical section, Communication source, Security token, Computer science, Node (physics), Set (abstract data type)

Related papers

Back to paper searchBrowse research topicsOriginal source
An improved √N algorithm for mutual exclusion in decentralized systems — Research Paper | ScholarLens