#### DMCA

## Fault Tolerant External Memory Algorithms

Citations: | 3 - 2 self |

### Citations

276 |
Organization and maintenance of large ordered indexes
- Bayer, McCreight
- 1972
(Show Context)
Citation Context ...res Sort(N) = Θ(N/B log M/B(N/B)) I/Os. See recent surveys [15,16] for an overview of other results. In the I/O model, a comparison based dictionary with optimal queries can be achieved with a B-tree =-=[17]-=-, which supports queries and updates in O(log B N) I/Os. Current resilient algorithms do not scale past the internal memory of a computer and thus, it is currently not possible to work with large sets... |

244 | Algorithm-based fault tolerance for matrix operations - Huang, Abraham - 1984 |

184 | Fault-Tolerant Computer System Design - Pradhan - 1996 |

83 | Algorithms and data structures for external memory
- Vitter
(Show Context)
Citation Context ...was presented in [14] For the I/O model, a comprehensive list of results have been achieved. It is shown in [8] that sorting N elements requires Sort(N) = Θ(N/B log M/B(N/B)) I/Os. See recent surveys =-=[15,16]-=- for an overview of other results. In the I/O model, a comparison based dictionary with optimal queries can be achieved with a B-tree [17], which supports queries and updates in O(log B N) I/Os. Curre... |

73 | External memory data structures
- Arge
- 2001
(Show Context)
Citation Context ...was presented in [14] For the I/O model, a comprehensive list of results have been achieved. It is shown in [8] that sorting N elements requires Sort(N) = Θ(N/B log M/B(N/B)) I/Os. See recent surveys =-=[15,16]-=- for an overview of other results. In the I/O model, a comparison based dictionary with optimal queries can be achieved with a B-tree [17], which supports queries and updates in O(log B N) I/Os. Curre... |

48 | MJRTY - A fast majority vote algorithm
- Boyer, Moore
- 1982
(Show Context)
Citation Context ...cells. Since at most δ of the copies can be corrupted, the majority of the 2δ + 1 elements are uncorrupted. The value can be retrieved using O( δ B ) I/Os and O(δ) time with the majority algorithm in =-=[18]-=-, which scans the 2+ . 1 values keeping a single majority candidate and a counter in reliable memory. A sequence is faithfully ordered if the uncorrupted elements form a sorted subsequence. 2 Lower Bo... |

45 | An approach to concurrent control flow checking - Yau, Chen - 1980 |

35 | Worst-case efficient external-memory priority queues
- Brodal, Katajainen
- 1998
(Show Context)
Citation Context ...nd a delete-min deletes and returns the minimum uncorrupted element or a corrupted one. Our priority queue is based on an amortized version of the worst-case optimal external memory priority queue of =-=[20]-=- using our new resilient multi-way merging algorithm to move elements between disk and internal memory. Details will appear in the full paper. Theorem 5. There is a linear space resilient priority que... |

21 | Sorting and searching in the presence of memory faults (without redundancy
- Finocchi, Italiano
(Show Context)
Citation Context ... corruptions. Corruptions can also often be discovered by existing hardware techniques, but even these techniques can fail and let some corrupted data take part of computations. Finocchi and Italiano =-=[7]-=- introduced the faulty-memory random access machine, based on the traditional RAM model. In this model, memory corruptions can occur at any time and at any place in memory during the execution of an a... |

20 |
J.S.: The input/output complexity of sorting and related problems
- Aggarwal, Vitter
- 1988
(Show Context)
Citation Context ...high degree of locality in their memory access pattern, that is, algorithms where data accessed close in time is also stored close in memory. This situation is modeled in the I/O model of computation =-=[8]-=-. In this model a disk of unlimited size and a memory of size M are available. Elements are transferred between disk and memory in blocks of size B and computation is performed on elements in memory o... |

17 | Optimal resilient sorting and searching in the presence of memory faults
- FINOCCHI, GRANDONI, et al.
- 2009
(Show Context)
Citation Context ...mplexity measure is the number of block transfers (I/Os) performed. Previous Work: Several problems have been addressed in the faulty-memory RAM, see a very recent survey [9] for more information. In =-=[10,7]-=-, matching upper and lower bounds for resilient sorting and randomized searching were given. Sorting N elements requires Θ(N log N+αδ) time [7]. Searching in a sorted array requires Ω(log N+δ) time, a... |

16 | Experimental evaluation of the fail-silent behaviour in programs with consistency checks - Rela, Madeira, et al. - 1996 |

15 | Resilient search trees
- Finocchi, Grandoni, et al.
(Show Context)
Citation Context ... Ω(log N+δ) time, and an optimal deterministic algorithm matching that bound is described in [11]. It has been empirically shown that resilient algorithms are of practical interest [12]. Recently, in =-=[11,13,14]-=- resilient data structuresI/O Complexity Assumptions I/O Tolerance Time Tolerance (max δ) (max δ) Det. Dict. O ` 1 ε logB N + δ B1−ε ´ 1 log B < ε < 1 O(B1−ε logB N) O(log N) Ran. Dict. O(logB N + δ ... |

14 |
A.: "Testing semiconductor memories - Theory and Practice
- Goor
(Show Context)
Citation Context ...paper we conduct the first study of algorithms and data structures for external memory in the presence of an unreliable internal and external memory. Contemporary memory devices such as SRAM and DRAM =-=[1, 2]-=- can be unreliable due to a number of factors, such as power failures, radiation, and cosmic rays. The content of a cell in unreliable memory can be silently altered and in standard memory circuits th... |

13 | Priority queues resilient to memory faults
- Jørgensen, Moruz, et al.
(Show Context)
Citation Context ... Ω(log N+δ) time, and an optimal deterministic algorithm matching that bound is described in [11]. It has been empirically shown that resilient algorithms are of practical interest [12]. Recently, in =-=[11,13,14]-=- resilient data structuresI/O Complexity Assumptions I/O Tolerance Time Tolerance (max δ) (max δ) Det. Dict. O ` 1 ε logB N + δ B1−ε ´ 1 log B < ε < 1 O(B1−ε logB N) O(log N) Ran. Dict. O(logB N + δ ... |

12 | Optimal resilient dynamic dictionaries
- Brodal, Fagerberg, et al.
(Show Context)
Citation Context ...nd to minimize the expression for α = δ, this bound matches the lower bound. Thus, this dictionary is optimal. Our data structure is based on the B-tree and the resilient binary search algorithm from =-=[11]-=-. In a standard B-tree search one corrupted element can misguide the algorithm, forcing at least one I/O in the wrong part of the tree. To circumvent this problem, each guiding element in each interna... |

11 | Designing reliable algorithms in unreliable memories
- FINOCCHI, GRANDONI, et al.
- 2005
(Show Context)
Citation Context ...ements in memory only. The complexity measure is the number of block transfers (I/Os) performed. Previous Work: Several problems have been addressed in the faulty-memory RAM, see a very recent survey =-=[9]-=- for more information. In [10,7], matching upper and lower bounds for resilient sorting and randomized searching were given. Sorting N elements requires Θ(N log N+αδ) time [7]. Searching in a sorted a... |

9 |
Cache-oblivious search trees via binary trees of small height
- Brodal, Fagerberg, et al.
- 2002
(Show Context)
Citation Context ...mma 2 and 3 by setting c = 1 and c ∈ [ 1 log B , 1 − log log B log B ≤ c ≤ 1, queries use log B , c = 1 − log log B log B , 1 ] such that c logB N = δ B1−c respectively. By adapting the techniques of =-=[11,19]-=- and the static dictionary presented above we obtain a dynamic dictionary. Details will appear in the full paper. Theorem 3. There is a deterministic dynamic resilient dictionary supporting searches a... |

8 | The Price of Resiliency: a Case Study on Sorting with Memory Faults - Petrillo, Finocchi, et al. - 2006 |