## Improved algorithms for dynamic page migration (2005)

Venue: | In Proc. of the 22nd Symp. on Theoretical Aspects of Computer Science (STACS |

Citations: | 6 - 5 self |

### BibTeX

@INPROCEEDINGS{Bienkowski05improvedalgorithms,

author = {Marcin Bienkowski and Miroslaw Dynia and Miroslaw Korzeniowski},

title = {Improved algorithms for dynamic page migration},

booktitle = {In Proc. of the 22nd Symp. on Theoretical Aspects of Computer Science (STACS},

year = {2005},

pages = {365--376}

}

### OpenURL

### Abstract

Abstract The dynamic page migration problem [4] is defined in a distributed network of n mobile nodes sharing one indivisible memory page of size D. During runtime, the nodes can both access a unit of data from the page and move with a constant speed, thus changing the costs of communication. The problem is to compute online a schedule of page movements to minimize the total communication cost. In this paper we construct and analyze the first deterministic algorithm for this problem. We prove that it achieves an (up to a constant factor) optimal competitive ratio O(n · √ D). We show that the randomization of this algorithm improves this ratio to O ( √ D · log n) (against an oblivious adversary). This substantially improves an O(n · √ D) upper bound from [4]. We also give an almost matching lower bound of Ω ( √ D · √ log n) for this problem. 1

### Citations

729 |
Tarjan. Amortized efficiency of list update and paging rules
- Sleator, E
- 1985
(Show Context)
Citation Context ...decide, online, when and where to move the page to minimize the total cost of communication over all sequences of requests. The performance of the online algorithm is measured by competitive analysis =-=[9, 6]-=-, i.e. by comparing its total cost to the total cost of the optimal offline algorithm on the same input sequence. The dynamic page migration (DPM) problem introduced by Bienkowski, Korzeniowski and Me... |

649 |
Online computation and competitive analysis
- Borodin, El-Yaniv
- 1998
(Show Context)
Citation Context ...decide, online, when and where to move the page to minimize the total cost of communication over all sequences of requests. The performance of the online algorithm is measured by competitive analysis =-=[9, 6]-=-, i.e. by comparing its total cost to the total cost of the optimal offline algorithm on the same input sequence. The dynamic page migration (DPM) problem introduced by Bienkowski, Korzeniowski and Me... |

164 | Competitive paging algorithms
- Fiat, Karp, et al.
- 1991
(Show Context)
Citation Context ...nted in this paper use the marking technique, which bears a resemblance to the Least Recently Used (LRU) paging algorithm by Sleator and Tarjan [9] and randomized MARK paging algorithm by Fiat et al. =-=[8]-=-, respectively. Contribution of the Paper. In this paper we improve and extend the results of [4]. In Sect. 3 we give a first deterministic algorithm, MARK, for the DPM problem. Our algorithm achieves... |

138 | On the power of randomization in on-line algorithms
- Ben-David, Borodin, et al.
- 1994
(Show Context)
Citation Context ... expected cost is not greater than c times the cost of the optimum. The expected value is taken over all possible random choices of the algorithm. In this paper we consider only oblivious adversaries =-=[3]-=-, which have no access to the random bits used by the algorithm. Related Work. For the page migration problem in a static network, Westbrook [10] gave the first randomized O(1)-competitive algorithms ... |

106 | Competitive Distributed File Allocation
- Awerbuch, Bartal, et al.
- 1993
(Show Context)
Citation Context ...dversary). This substantially improves an O(n · √ D) upper bound from [4]. We also give an almost matching lower bound of Ω( √ D · √ log n) for this problem. 1 Introduction The page migration problem =-=[1, 2, 5, 7, 10]-=- arises in a distributed network of processors which share some global data. Shared variables or memory pages are stored at the local memory of these processors. If a processor wants to access (read o... |

62 | Competitive algorithms for replication and migration poblems - Black, Sleator - 1989 |

29 | Randomized algorithms for multiprocessor page migration
- Westbrook
- 1992
(Show Context)
Citation Context ...dversary). This substantially improves an O(n · √ D) upper bound from [4]. We also give an almost matching lower bound of Ω( √ D · √ log n) for this problem. 1 Introduction The page migration problem =-=[1, 2, 5, 7, 10]-=- arises in a distributed network of processors which share some global data. Shared variables or memory pages are stored at the local memory of these processors. If a processor wants to access (read o... |

28 | On page migration and other relaxed task systems
- Bartal, Charikar, et al.
(Show Context)
Citation Context ...dversary). This substantially improves an O(n · √ D) upper bound from [4]. We also give an almost matching lower bound of Ω( √ D · √ log n) for this problem. 1 Introduction The page migration problem =-=[1, 2, 5, 7, 10]-=- arises in a distributed network of processors which share some global data. Shared variables or memory pages are stored at the local memory of these processors. If a processor wants to access (read o... |

13 | Page migration algorithms using work functions
- Chrobak, Larmore, et al.
- 1992
(Show Context)
Citation Context |

5 |
auf der Heide, Fighting against two adversaries: Page migration in dynamic networks
- Bienkowski, Korzeniowski, et al.
- 2004
(Show Context)
Citation Context ...of Paderborn, Germany {young, rudy}@upb.de 2 DFG Graduate College “Automatic Configuration in Open Systems”, University of Paderborn, Germany mdynia@upb.de Abstract The dynamic page migration problem =-=[4]-=- is defined in a distributed network of n mobile nodes sharing one indivisible memory page of size D. During runtime, the nodes can both access a unit of data from the page and move with a constant sp... |

4 |
Probabilistic computation: towards a uniform measure of complexity, in
- Yao
- 1977
(Show Context)
Citation Context ...tribution over inputs and prove that no deterministic algorithm, which knows this distribution, can have a competitive ratio better than Ω(min{ √ D · log n, λ}). Then one can apply Yao minmax theorem =-=[11, 6]-=- and Thm. 3 follows immediately. We assume that n is a power of 2. If it is not the case, then the adversary can give requests only in the first 2 ⌊log n⌋ nodes and put the other nodes exactly in the ... |