Another solution of the mutual exclusion problem
Information Processing Letters
Algorithms for mutual exclusion
Algorithms for mutual exclusion
A fast mutual exclusion algorithm
ACM Transactions on Computer Systems (TOCS)
Proof of a mutual exclusion algorithm—a classic example
ACM SIGOPS Operating Systems Review
Concurrent programming
A bounded first-in, first-enabled solution to the l-exclusion problem
ACM Transactions on Programming Languages and Systems (TOPLAS)
Using k-exclusion to implement resilient, scalable shared objects (extended abstract)
PODC '94 Proceedings of the thirteenth annual ACM symposium on Principles of distributed computing
Bounded Concurrent Time-Stamping
SIAM Journal on Computing
On concurrent programming
Peterson's mutual exclusion algorithm revisited
Science of Computer Programming
Solution of a problem in concurrent programming control
Communications of the ACM
Distributed Algorithms
Deriving a Scalable Algorithm for Mutual Exclusion
DISC '98 Proceedings of the 12th International Symposium on Distributed Computing
Adaptive solutions to the mutual exclusion problem
Distributed Computing
Resource allocation with immunity to limited process failure
SFCS '79 Proceedings of the 20th Annual Symposium on Foundations of Computer Science
Hi-index | 0.00 |
Peterson's n-process mutual exclusion algorithm [P81] has been widely touted for elegance and simplicity. It has been analyzed extensively, and yet certain properties have eluded the researchers. This paper illustrates, and expands on, several properties of Peterson's algorithm: (1) We reiterate that the number of processes that can overtake a process, called unfairness index, is unbounded in Peterson's algorithm; (2) With a slight modification of the algorithm, we obtain the unfairness index of n(n - 1)/2; (3) We identify an inherent characteristic of that algorithm that sets the lower bound of n(n - 1)/2 for the unfairness index; (4) By modifying the characteristic, we obtain algorithms with unfairness index (n - 1); (5) We show that the new algorithms are amenable to reducing shared space requirement, and to improving time efficiency (where the number of steps executed is proportional to the current contention); and (6) We also extend the algorithms to solve l-exclusion problem in a simple and straightforward way.