• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Efficient allocation in distributed object oriented databases with capacity and security constraints.
【24h】

Efficient allocation in distributed object oriented databases with capacity and security constraints.

机译：具有容量和安全性约束的分布式面向对象数据库中的有效分配。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Efficient distribution of data is a major challenge in distributed databases. The problem is even more severe for distributed object oriented databases because of inheritance, encapsulation and the more complex problem involved when methods invoke other methods. This problem, the Object Allocation Problem (OAP), is a harder version of the relational database allocation problem (DAP), a problem known to be NP-hard. The additional problem of restricting sensitive data to secure sites, is at least as hard as the OAP.; In this research we investigated the cost efficient distribution of a fragmented object oriented database to the sites of a network. We investigated allocation algorithms for distributing the fragments in the presence of both capacity and security constraints. Initially, we assumed that all fragments were sensitive and all sites secure, and concentrated on minimizing the number of external accesses in the presence of only the capacity constraints at each node. We later relaxed that restriction and performed the optimal allocation assuming that a percentage of the fragments were sensitive and a percentage of the sites were secure. We also looked at placing the frequently communicating fragments in close proximity. The two proximity placement problems we looked at, were the adjacency placement and minimum diameter placement of frequently communicating fragments. We showed these two problems to be NP-complete.; Of the three algorithms we implemented (Genetic algorithm (GA), Simulated annealing (SA) and Kernighan-Lin (KL)) the SA and GA algorithms were clearly superior. In the presence of both capacity and security constraints, the various versions of the GA performed consistently better in minimizing the number of security violations while the SA algorithm versions typically had lower costs. The noted exception was the GA penalty version algorithm with random initial allocation, which performed very well in both cost minimization and reducing the number of security violations.
机译：高效的数据分发是分布式数据库中的主要挑战。对于分布式面向对象的数据库，由于继承，封装以及方法调用其他方法时涉及的更为复杂的问题，该问题甚至更加严重。这个问题，即对象分配问题（OAP），是关系数据库分配问题（DAP）的较难版本，该问题被称为NP难题。将敏感数据限制到安全站点的附加问题至少与OAP一样困难。在这项研究中，我们调查了面向对象的碎片数据库向网络站点的经济有效分配。我们研究了在容量和安全性约束均存在的情况下用于分配碎片的分配算法。最初，我们假设所有片段都是敏感的，所有站点都是安全的，并且集中精力在每个节点上仅存在容量限制的情况下将外部访问的数量减至最少。后来我们放宽了限制，并假设一定百分比的片段是敏感的并且一定比例的站点是安全的，并进行了最佳分配。我们还研究了将经常交流的片段放在附近。我们研究的两个邻近放置问题是频繁通信的片段的邻近放置和最小直径放置。我们证明这两个问题是NP完全的。在我们实施的三种算法（遗传算法（GA），模拟退火（SA）和克尼根·林（KL））中，SA和GA算法明显优越。在容量和安全性约束同时存在的情况下，GA的各种版本在减少安全违规的次数方面始终表现出更好的性能，而SA算法版本通常具有较低的成本。值得注意的例外是具有随机初始分配的GA惩罚版本算法，该算法在最小化成本和减少违反安全性方面均表现出色。

著录项

  • 作者

    Graham, Jonathan.;

  • 作者单位

    University of Idaho.;

  • 授予单位 University of Idaho.;
  • 学科 Computer Science.
  • 学位 Ph.D.
  • 年度 2005
  • 页码 139 p.
  • 总页数 139
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 自动化技术、计算机技术;
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号