基于本地化差分隐私多关系表示上的Star-JOIN查询已得到研究者广泛关注.现有基于OLH机制与层次树结构的Star-JOIN查询算法存在根节点泄露隐私风险、τ-截断机制没有给出如何选择合适τ值等问题.针对现有算法存在的不足,提出一种有效且...基于本地化差分隐私多关系表示上的Star-JOIN查询已得到研究者广泛关注.现有基于OLH机制与层次树结构的Star-JOIN查询算法存在根节点泄露隐私风险、τ-截断机制没有给出如何选择合适τ值等问题.针对现有算法存在的不足,提出一种有效且满足本地化差分隐私的Star-JOIN查询算法LPRR-JOIN(longitudinal path random response for join).该算法充分利用层次树的纵向路径结构与GRR机制,设计一种纵向本地扰动算法LPRR,该算法以所有属性纵向路径上的节点组合作为扰动值域.每个用户把自身元组映射到相应节点组合中,再利用GRR机制对映射后的元组进行本地扰动.为了避免事实表上存在的频率攻击,LPRR-JOIN算法允许每个用户利用阈值τ本地截断自身元组个数,大于τ条元组删减、小于τ条元组补充.为了寻找合适的τ值,LPRR-JOIN算法利用τ-截断带来的偏差与扰动方差构造总体误差函数,通过优化误差目标函数获得τ值;其次结合用户分组策略获得τ值的总体分布,再利用中位数获得合适的τ值.LPRR-JOIN算法与现有算法在3种多关系数据集上进行比较,实验结果表明其响应查询算法优于同类算法.展开更多
This paper proposes a semi-greedy framework for optimizing multi-joinqueries in shared-nothing systems. The plan generated by the framework com-prises several pipelines, each performing several joins. The framework de...This paper proposes a semi-greedy framework for optimizing multi-joinqueries in shared-nothing systems. The plan generated by the framework com-prises several pipelines, each performing several joins. The framework deter-mines the 'optimal' number of joins to be performed in each pipeline. Thedecisions are made based on the cost estimation of the entire processing plan.Two ekisting optimization algorithms are extended under the framework. Ananalytical model is presented and used to compare the quality of plans producedby each optimization algorithm. Our study shows that the new algorithms out-perform their counterparts that are not extended.展开更多
文摘基于本地化差分隐私多关系表示上的Star-JOIN查询已得到研究者广泛关注.现有基于OLH机制与层次树结构的Star-JOIN查询算法存在根节点泄露隐私风险、τ-截断机制没有给出如何选择合适τ值等问题.针对现有算法存在的不足,提出一种有效且满足本地化差分隐私的Star-JOIN查询算法LPRR-JOIN(longitudinal path random response for join).该算法充分利用层次树的纵向路径结构与GRR机制,设计一种纵向本地扰动算法LPRR,该算法以所有属性纵向路径上的节点组合作为扰动值域.每个用户把自身元组映射到相应节点组合中,再利用GRR机制对映射后的元组进行本地扰动.为了避免事实表上存在的频率攻击,LPRR-JOIN算法允许每个用户利用阈值τ本地截断自身元组个数,大于τ条元组删减、小于τ条元组补充.为了寻找合适的τ值,LPRR-JOIN算法利用τ-截断带来的偏差与扰动方差构造总体误差函数,通过优化误差目标函数获得τ值;其次结合用户分组策略获得τ值的总体分布,再利用中位数获得合适的τ值.LPRR-JOIN算法与现有算法在3种多关系数据集上进行比较,实验结果表明其响应查询算法优于同类算法.
文摘This paper proposes a semi-greedy framework for optimizing multi-joinqueries in shared-nothing systems. The plan generated by the framework com-prises several pipelines, each performing several joins. The framework deter-mines the 'optimal' number of joins to be performed in each pipeline. Thedecisions are made based on the cost estimation of the entire processing plan.Two ekisting optimization algorithms are extended under the framework. Ananalytical model is presented and used to compare the quality of plans producedby each optimization algorithm. Our study shows that the new algorithms out-perform their counterparts that are not extended.
文摘联机分析处理OLAP(online analytical processing)查询作为一种复杂查询,当使用SQL(structured query language)语句来表述时,通常都包含多表连接和分组聚集操作,因此提高多表连接和分组聚集计算的性能就成为ROLAP(relational OLAP)查询处理的关键问题.提出一种基于分组序号的聚集算法MuGA(group number based aggregation with multi-table join),该方法充分考虑数据仓库星型模式的特点,将聚集操作和新的多表连接算法MJoin(multi-table join)相结合,使用分组序号进行分组聚集计算,代替通常的排序或者哈希计算,从而有效地减少CPU运算以及磁盘存取的开销.算法的实验数据表明,提出的MuGA算法与传统的关系数据库聚集查询处理方法以及改进后的基于排序的聚集算法相比,性能都有显著提高.
