### Abstract

First Come First Served (FCFS) is a policy that is accepted for implementing fairness in a number of application domains such as scheduling in Operating Systems, scheduling web requests, and so on. We also have orthogonal applications of FCFS policies in proving correctness of search algorithms such as Breadth-First Search, the Bellman-Ford FIFO implementation for finding single-source shortest paths, program verification and static analysis. The data structure used to implementing FCFS policies, the queue, suffers from two principal drawbacks, viz., non-trivial verifiability and lack of scalability. In case of large distributed networks, maintaining an explicit queue to enforce FCFS is prohibitively expensive. The question of interest then, is whether queues axe required to implement FCFS policies; this paper provides empirical evidence answering this question in the negative. The principal contribution of this paper is the design and analysis of a randomized protocol to implement approximate FCFS policies without queues. From the Software Engineering perspective, the techniques that are developed find direct applications in program verification, model checking, in the implementation of distributed queues and in the design of incremental algorithms for Shortest path problems.

Original language | English (US) |
---|---|

Title of host publication | High Performance Computing - HiPC 2007 - 14th International Conference, Proceedings |

Pages | 540-551 |

Number of pages | 12 |

State | Published - Dec 1 2007 |

Event | 14th International Conference on High-Performance Computing, HiPC 2007 - Goa, India Duration: Dec 18 2007 → Dec 21 2007 |

### Publication series

Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|

Volume | 4873 LNCS |

ISSN (Print) | 0302-9743 |

ISSN (Electronic) | 1611-3349 |

### Other

Other | 14th International Conference on High-Performance Computing, HiPC 2007 |
---|---|

Country | India |

City | Goa |

Period | 12/18/07 → 12/21/07 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Theoretical Computer Science
- Computer Science(all)

### Cite this

*High Performance Computing - HiPC 2007 - 14th International Conference, Proceedings*(pp. 540-551). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4873 LNCS).

}

*High Performance Computing - HiPC 2007 - 14th International Conference, Proceedings.*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4873 LNCS, pp. 540-551, 14th International Conference on High-Performance Computing, HiPC 2007, Goa, India, 12/18/07.

**Accomplishing approximate FCFS fairness without queues.** / Subramani, K.; Madduri, Kamesh.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Accomplishing approximate FCFS fairness without queues

AU - Subramani, K.

AU - Madduri, Kamesh

PY - 2007/12/1

Y1 - 2007/12/1

N2 - First Come First Served (FCFS) is a policy that is accepted for implementing fairness in a number of application domains such as scheduling in Operating Systems, scheduling web requests, and so on. We also have orthogonal applications of FCFS policies in proving correctness of search algorithms such as Breadth-First Search, the Bellman-Ford FIFO implementation for finding single-source shortest paths, program verification and static analysis. The data structure used to implementing FCFS policies, the queue, suffers from two principal drawbacks, viz., non-trivial verifiability and lack of scalability. In case of large distributed networks, maintaining an explicit queue to enforce FCFS is prohibitively expensive. The question of interest then, is whether queues axe required to implement FCFS policies; this paper provides empirical evidence answering this question in the negative. The principal contribution of this paper is the design and analysis of a randomized protocol to implement approximate FCFS policies without queues. From the Software Engineering perspective, the techniques that are developed find direct applications in program verification, model checking, in the implementation of distributed queues and in the design of incremental algorithms for Shortest path problems.

AB - First Come First Served (FCFS) is a policy that is accepted for implementing fairness in a number of application domains such as scheduling in Operating Systems, scheduling web requests, and so on. We also have orthogonal applications of FCFS policies in proving correctness of search algorithms such as Breadth-First Search, the Bellman-Ford FIFO implementation for finding single-source shortest paths, program verification and static analysis. The data structure used to implementing FCFS policies, the queue, suffers from two principal drawbacks, viz., non-trivial verifiability and lack of scalability. In case of large distributed networks, maintaining an explicit queue to enforce FCFS is prohibitively expensive. The question of interest then, is whether queues axe required to implement FCFS policies; this paper provides empirical evidence answering this question in the negative. The principal contribution of this paper is the design and analysis of a randomized protocol to implement approximate FCFS policies without queues. From the Software Engineering perspective, the techniques that are developed find direct applications in program verification, model checking, in the implementation of distributed queues and in the design of incremental algorithms for Shortest path problems.

UR - http://www.scopus.com/inward/record.url?scp=38349048706&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38349048706&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:38349048706

SN - 9783540772194

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 540

EP - 551

BT - High Performance Computing - HiPC 2007 - 14th International Conference, Proceedings

ER -