面向查询的索引结构 操作方法 口初始 每个对象计算其所在的查询。 口增加新的查询 ■放入索引结构 计算包含的对象 口时间变化 监测移动对象的位置变化 ■计算查询结果的变化
11 面向查询的索引结构 ◼ 操作方法 ❑ 初始 ◼ 每个对象计算其所在的查询。 ❑ 增加新的查询 ◼ 放入索引结构 ◼ 计算包含的对象 ❑ 时间变化 ◼ 监测移动对象的位置变化 ◼ 计算查询结果的变化
Query view a query view is a combined query consisting of one or more continuous queries q1, . qi. The monitoring region of the query view, called the view region, is defined as the intersection of the monitoring regions of q1, . qi a In the Figure, there are five query views v1 to v5, which correspond to q1如q1,q2}如q2.{q2q3}and{q3} V5 乎 V4 V3 V2 Vl Fig 1. An example of query views
12 Query View ◼ A query view is a combined query consisting of one or more continuous queries q1, ..., qi. The monitoring region of the query view, called the view region, is defined as the intersection of the monitoring regions of q1, ..., qi ❑ In the Figure, there are five query views v1 to v5, which correspond to {q1}, {q1, q2}, {q2}, {q2,q3} and {q3}
Query view For the construction of query views, we take splitting- merging"strategy a Divide the space into small grid cells of same size and map all the monitoring regions on the grid space Then we collect all the arid cells, and merge those that are covered by the same query regions between the objects'movements and the 3 Query view abstracts the dynamic relati query answers a The query answers are changed if and only if there are objects changing their query views o CEM incurs the minimum transmission cost on the moving object side
13 Query View ◼ For the construction of query views, we take “splitting-merging” strategy ❑ Divide the space into small grid cells of same size and map all the monitoring regions on the grid space. Then we collect all the grid cells, and merge those that are covered by the same query regions. ◼ Query view abstracts the dynamic relation between the objects’ movements and the query answers ❑ The query answers are changed if and only if there are objects changing their query views ❑ CEM incurs the minimum transmission cost on the moving object side
Data structure On the server side a Query view Table(QvT) To store the query view id its view region and the corresponding query list a Grid-based View Index (GVI A grid structure to index the query views a Query Table(QT) To organize the queries a Query Result Buffer(QRB) To store the incoming results 14
14 Data Structure ◼ On the server side ❑ Query View Table (QVT) ◼ To store the query view id, its view region, and the corresponding query list ❑ Grid-based View Index (GVI) ◼ A grid structure to index the query views ❑ Query Table (QT) ◼ To organize the queries ❑ Query Result Buffer (QRB) ◼ To store the incoming results
Data structure On the mobile side a Query View Region (QVR To represent the moving object's current view region a Grid-based Location IndeX GLI An auxiliary structure to fast determine whether the current location is within the view region
15 Data Structure ◼ On the mobile side ❑ Query View Region (QVR) ◼ To represent the moving object’s current view region ❑ Grid-based Location Index (GLI) ◼ An auxiliary structure to fast determine whether the current location is within the view region