《Oracle数据库11g在OracleSpatial网络数据模型中处理大型网络的一种按需加载方法资料》由会员分享,可在线阅读,更多相关《Oracle数据库11g在OracleSpatial网络数据模型中处理大型网络的一种按需加载方法资料(27页珍藏版)》请在金锄头文库上搜索。
1、 An Oracle White Paper February 2009 A Load On Demand Approach to Handling Large Networks in the Oracle Spatial Network Data Model Oracle White Paper A Load On Demand Approach to Handling Large Networks in the Oracle Spatial Network Data Model Disclaimer The following is intended to outline our gene
2、ral product direction It is intended for information purposes only and may not be incorporated into any contract It is not a commitment to deliver any material code or functionality and should not be relied upon in making purchasing decisions The development release and timing of any features or fun
3、ctionality described for Oracle s products remains at the sole discretion of Oracle Oracle White Paper A Load On Demand Approach to Handling Large Networks in the Oracle Spatial Network Data Model Executive Overview 2 Introduction 3 Oracle Spatial Network Data Model 5 Network Data Model Schema 5 Net
4、work Partitioning 7 NDM LOD Architecture and APIs 8 LOD Network Analysis Capabilities 8 Java Representations of LOD Network Elements 9 Software Requirements 9 Using LOD in Network Data Model 9 Creating a Network 10 Partitioning a Network 10 Configuring the Partition Cache 11 Analyzing the Network 12
5、 Modeling and Analysis Enhancements 15 Network Constraints 15 Multiple Cost Support 16 Precomputed Connected Components 17 Dynamic Data Set 18 Partial Link Paths Sub paths 19 Hierarchical Shortest Path Computation 19 Comparisons between NDM in memory API and NDM LOD API 21 Approaches 21 Data Model 2
6、1 Analysis Functions 21 LOD Analysis Viewer 23 Conclusion 24 References 24 Oracle White Paper A Load On Demand Approach to Handling Large Networks in the Oracle Spatial Network Data Model 2 Executive Overview Network modeling management and analysis are common tasks for enterprise applications such
7、as geographic information system GIS customer relationship management CRM social network analysis and in semantic web technologies such as the resource description framework RDF Oracle10g introduced the Oracle Spatial network data model NDM which lets users model and analyze networks In Oracle10g ND
8、M uses an in memory approach to pre load the whole network into memory before analysis however this approach cannot handle networks that are too large to fit in memory To address this scalability issue we developed a load on demand approach LOD in Oracle11g Large networks are first divided into mana
9、geable parts called network partitions Only partitions that are needed are automatically loaded during analysis In this paper we discuss how LOD works and how to use it Oracle White Paper A Load On Demand Approach to Handling Large Networks in the Oracle Spatial Network Data Model 3 Introduction The
10、 network data model helps users analyze network connectivity relationships It is commonly used in transportation utilities life sciences and semantic technologies It simplifies network modeling analysis and management so that users can focus on application logic It provides an open generic data mode
11、l with many common network analysis capabilities Application information is separated from connectivity information so that the model can be applied to many network applications without customization NDM further provides a constraint mechanism to let users guide analysis based on application rules a
12、nd attributes For more information on the Oracle Spatial Network Data Model see references 1 2 3 On Oracle10g NDM APIs use an in memory approach to analyzing networks The entire network is loaded into memory from the database Once the network is loaded in memory users can query and edit it This appr
13、oach works well for networks that can be completely loaded into memory however it cannot handle networks that are too big to be fit in memory To address this scalability issue Oracle11g introduces a load on demand approach to handle large networks Instead of loading the whole network into memory the
14、 network is first divided into manageable subnetworks network partitions and only partitions that are needed during analysis are loaded into memory Loading and unloading of network partitions are automatically managed thus removing memory as a limiting factor Figure 1 shows a US major highway road n
15、etwork The complete US road network contains around 20 million nodes and 50 millions links source Navteq 2006 Networks of this size cannot be handled in memory and need to be handled by a load on demand approach Oracle White Paper A Load On Demand Approach to Handling Large Networks in the Oracle Sp
16、atial Network Data Model 4 Figure 1 US road network 20 million nodes and 50 millions links source Navteq LOD uses the same network data model in database as the in memory approach network node link and path tables and the network metadata In addition though LOD requires networks to be partitioned first NDM provides a partitioning procedure for spatial networks which enables users to partition their spatial networks into network partitions and store the result in a partition table To further spee