公开(公告)号：WO2013002870A3

公开(公告)日：2014-05-01

申请号：PCT/US2012032891

申请日：2012-04-10

Applicant: IBM ,  CHANG YUAN-CHI ,  KOVED LAWRENCE ,  LANG CHRISTIAN A ,  SOMINSKY IGOR

Inventor： CHANG YUAN-CHI ,  KOVED LAWRENCE ,  LANG CHRISTIAN A ,  SOMINSKY IGOR

IPC: G06F17/30

CPC classification number: G06F17/30551 ,  G06F17/30489

Abstract: Techniques are disclosed for multi-granularity hierarchical aggregate selection based on update, storage and response constraints. For example, for a temporal hierarchy of aggregation statistics associated with a plurality of database records, wherein the temporal hierarchy comprises two or more aggregation statistics levels and each level has a different temporal granularity associated therewith, a method comprises iteratively modifying the temporal hierarchy to at least one of: (a) minimize a storage usage cost while satisfying a temporal hierarchy update constraint and a query response time constraint; (b) reduce a temporal hierarchy update time and a query response time while satisfying a storage usage constraint; and (c) minimize a query response time for frequently applied queries that do not shift in time while satisfying the storage usage constraint, wherein the resulting temporal hierarchy that achieves at least one of (a), (b) and (c) is identified as an optimal temporal hierarchy.

Abstract translation: 公开了基于更新，存储和响应约束的多粒度分级聚合选择的技术。 例如，对于与多个数据库记录相关联的聚合统计的时间层级，其中时间层级包括两个或多个聚合统计级别，并且每个级别具有与之相关联的不同时间粒度，一种方法包括将时间层级迭代地修改为 至少一个：（a）在满足时间层次更新约束和查询响应时间约束的同时最小化存储使用成本; （b）在满足存储使用限制的同时减少时间层次更新时间和查询响应时间; （c）最小化在满足存储使用限制的同时不随时间移动的频繁应用的查询的查询响应时间，其中识别达到（a），（b）和（c）中的至少一个的所得到的时间层级， 作为最优时间层级。

公开(公告)号：GB2506800A

公开(公告)日：2014-04-09

申请号：GB201400973

申请日：2012-04-10

Applicant: IBM

Inventor： CHANG YUAN-CHI ,  KOVED LAWRENCE ,  LANG CHRISTIAN A ,  SOMINSKY IGOR

IPC: G06F17/30

Abstract: Techniques are disclosed for multi-granularity hierarchical aggregate selection based on update, storage and response constraints. For example, for a temporal hierarchy of aggregation statistics associated with a plurality of database records, wherein the temporal hierarchy comprises two or more aggregation statistics levels and each level has a different temporal granularity associated therewith, a method comprises iteratively modifying the temporal hierarchy to at least one of: (a) minimize a storage usage cost while satisfying a temporal hierarchy update constraint and a query response time constraint; (b) reduce a temporal hierarchy update time and a query response time while satisfying a storage usage constraint; and (c) minimize a query response time for frequently applied queries that do not shift in time while satisfying the storage usage constraint, wherein the resulting temporal hierarchy that achieves at least one of (a), (b) and (c) is identified as an optimal temporal hierarchy.