2. Oracle10g Database: A Spatial VLDB Case Study Xavier R. Lopez Director, Server Technologies Oracle Corporation Session id: 40125

3. Overview  Role of Spatial Database Technology  Oracle Spatial & Locator  Benefits of Oracle 10g in large systems  What ’ s Coming in 10g  VLDB Case Study – UK Ordnance Survey

4. Why Use a Spatial Database?  Every Oracle DBMS is a Spatial database  Integrate location & Business data in RDBMS  Open Access - no proprietary types  High Performance – 64 bit  Integrity managed by DBMS  Scaleable: Supports Terabytes of Data  Scaleable: Supports 1000s of Users  Easy to Program – SQL, Java, XML  Security and Reliability  Short & Long Transaction Management  You already license it !!

5. 5 Early Spatial Systems: Hybrids  Attributes in database  Geometries in proprietary data files  Poor integration  No access to spatial data from IT Attributes Geometries GIS IT SQL blobs Attributes in database Geometries in database - but in proprietary binary format IT can access geometries via proprietaty interfaces only Poor integration

6. 6 Open Spatial Databases GIS eBusiness Geometries Attributes SQL Spatial is native DBMS type Attributes and geometries integrated in database Supported by all GIS Supported by eBusiness applications Spatial data queried using SQL, Java SQL GIS eBusiness

7. Design Tools XML Database Business Intelligence Mobile & Wireless JDeveloper Tools 3 rd party GIS Tools ERP Applications Oracle Stack is Spatially Enabled CRM Applications Spatial

8. 8 Differentiating Oracle eBusiness Suite with Location Capability Competition (SAP, Siebel, SAS)  No native location capability  Location requires 3 rd party tool and integration cost  Multiple data models  3 rd party proprietary BLOBs or Long Raws  Complex technical support Oracle eBusiness Suite  Integrated Location capability  No additional mapping tools or integration needed  Single data model  Native Spatial format  Out-of-the box feature & support

9. 9 Location is Integrated throughout Oracle Platform Weather Positioning Sensors Field Obs. CRM Business Intell Push Pull Telemetry Services Map Rendering Wireless LBS Spatial Components Clients Application Server 10g Oracle10g e-Business Applications Soap XML Location FeaturesOracle Stack Locator … Spatial … ERP

10. 10 Location Components Location-enabled Enterprise: Complementary Roles CRM/ERPBusiness Intelligence SQL XML/SOAP Oracle Spatial 10g

11. Integrating Spatial & eBusiness  Location adds significant value providing graphical view of supply chain, assets, customers, suppliers

12. 12 A Spatially-enabled Database

13. Generating Map Data from Oracle Spatial Tables Fisher Circle 85th St. Coop Court Road ROAD_ID 1 2 3 SURFACE Asphalt NAME Pine Cir. 2nd St. 3rd St. LANES 4 2 LOCATION

14. 14 How Spatial Data Is Stored in 10g? Data type Geographic coordinates

15. 15 Performing Location Query on Oracle 10g Find post offices within 3km to my office? Main Street 163 Island Park Dr. K1Y 2C3 + Station B K1Y 2C4 3 km + Station P K1Y 2C3 SQL> SELECT P.Post_Office_Name, P.Address 2> FROM Post_Offices P, 3> Address_Master A 4> WHERE 5> A.St_Address =‘163 Island Park Dr.’ 6> and A.City = ‘Ottawa’ 7> AND MDSYS.SDO_WITHIN_DISTANCE( 8> A.Location, P.Location, 9> ‘distance=3’) = ‘TRUE’;

16. 16 Oracle10g Spatial Capabilities  R-tree index  Native Spatial Type  Spatial Operators  Spatial Reference System  Long Transactions  Parallel Index & Query  Geodetic Support

17. Oracle10g Location Features Locator  Points, lines, polys  2D, 3D, 4D data  Spatial Operators – Within-distance – Spatial Relations  Coordinate Systems  Long Transactions Table Partitioning*  Object Replication*  Oracle10g Standard & Enterprise Spatial  All Locator features  Linear Referencing  Spatial Aggregates  Coordinate Transforms  Network Data Model  Topology Data Model  GeoRaster  Geocoder  Spatial Analytic Functions Oracle10g Enterprise Option Only * Available on Enterprise Edition Only

18. Extensible Database Framework Optimizer Query Engine Index Engine Type Manager Extensibility

19. 19 Securing Spatial Information Access control Privacy & integrity of data Comprehensive auditing Boundary a Infrastructure Building a Point c Boundary c Infrastructure D Point b Boundary bPoint a Building bInfra B Build DInfra C Building C            Network Security Privacy & integrity of communications  Authenticate User Security Data Security Points of Interest Buildings Infrastructure Boundaries

20. 20 9i R1 Performance Enhancements  R-tree index creation up to 20% faster than 8i  Partitioned local indexes provide additional performance gains for queries on large datasets, and concurrent queries and updates  Spatial aggregate functions speed retrieval of large sets of SDO_GEOMETRY objects  WITHIN_DISTANCE queries run up to 40% faster with R-tree indexes  VALIDATE_GEOMETRY function runs 200% faster depending on the complexity of the geometries

21. 21 9i R2 Performance Enhancements  R-tree indexes are created 50% faster than 9i  R-tree update performance increased significantly  Parallel index builds dramatically reduce R-tree index creation for large non-point spatial datasets  Spatial queries using secondary filters on geodetic data run up to 40% faster  Applications that execute WITHIN_DISTANCE queries on linear features perform 30% faster.

22. 22 Ten Spatial “ Tips of the Trade ”  Use R-Tree Spatial Indexes for your GIS applications!  Ensure USER_SDO_GEOM_METADATA extents are correct (for longitude/latitude, must be (-180,180) (-90,90)  Ensure tolerance is set correctly in USER_SDO_GEOM_METADATA.  Always Validate geometries using either: SDO_GEOM.VALIDATE_GEOMETRY_WITH_CONTEXT or SDO_GEOM.VALIDATE_LAYER_WITH_CONTEXT functionS  Spatial objects must use 4-digit notation for types  You can load spatial data through your GIS tool, Oracle SQL*Loader, or other 3 rd party tools (eg. Safe FME)  Consider binary access only when single geometry exceeds 1M ordinates  Use Realistic Generalized Layers  Spatial objects must use 4-digit notation for types  Use DBMS tracing to identify performance bottlenecks

23. Oracle Spatial 10g Enables Critical Grid Capabilities  Integrated Data Management  Seamless Datasets  Supports Complex Workflows  Easy to Manage & Program  High Performance  Supports Terabytes of Data  Supports 1000s of Users  Fast Transactions  Parallelism  Bulk Load Utilities  Open/Standards-based  Reliable  Scalability  Distributed/Replication  Partitioning  Version Management  History Management  Security  Multi-platform Support

24. 24 What the Analysts are Saying about Oracle Spatial … “ In repeated surveys, IDC has found that Oracle is used in an 80%-90% share of Spatial Information Management oriented database installations. ” IDC, December 2002

25. 25 The Dominant Spatial Database  National Mapping, Cadasters & Hydrographic Agencies – NIMA, USGS, US Army, Ordnance Survey (UK, IR, NI), Denmark, Sweden, The Netherlands, Poland, Australia  Transportation Management – California, Iowa, Florida, Maine, Maryland, Minnesota, New York, Oklahoma, Pennsylvania, Alabama, Alberta, London Rail, Netherlands Transport, Australia, Austrian Rail, German Rail  Telco & Wireless LBS – AT&T, Bell South, Cingular, DoCoMo, KDDI, Intrado, JPhone, Nextel, Sprint, T- Mobile, Telkom, Telenor, Telstra, Telus, Telia, Cellcom, Verizon, VIAG, Vodaphone, Wind  Utilities – Omaha Public Power, Reliant, US DoE, Western Power Corp, Severn Trent, Bejing Power, Czech Telem, Copenhagen Energy, Electrable, Gaz de France, Hydro-Quebec, Equitable Resources, Nova Naturgas, Sao Paulo Electric,  Local Authorities – New York City, Chicago, Los Angeles, San Jose, San Mateo, Washington DC, Cleveland, Detroit, Phoenix, Winnipeg, Vancouver, Edmonton, Stockholm…

26. 26 What’s coming?

27. Upcoming Spatial 10g Features  Network Data Model  Topology Data Model  GeoRaster  Geocoder  Spatial Analytic Functions  MapViewer

28. 28 10g: Network Data Model  Network Data Model – A data model to store network (graph) structure in the database – Explicitly stores and maintains connectivity of the network – Attributes at link and node level  Routing Engine – Street navigation for single or multiple destinations – Provide network analysis functionality in the database  Supports network solutions (Tracing & Routing) – Transportation and Transit Solutions – Field Service, Logistics – Location based Services and Telematics  VP: Bio-Info Pathways (Life Sciences)

29. 29 10g: Topology Data Model  New data model to store persistent topology – Easier to do data consistency checks in this model – Example: when the road moves, the property boundary automatically moves with it  Topology Data Model and Schema – Describes how different spatial features are related to each other – A land parcel shares the boundary with a road  10g continues to support transient topology – Topology computed on demand Conference Center Oracle Pkwy

30. 30 10g: GeoRaster  What is GeoRaster – A new data type to store raster data  Satellite images, remote sensing data – An XML schema to store Metadata  Data source, layer information – Geo Referencing information  How to relate a pixel in the image to a longitude/latitude on Earth’s surface  Functionality – Open, general purpose raster data model – storage and indexing of raster data  No size limit for each raster object – querying and analyzing raster data – delivering GeoRaster to external consumers  Publish as JPEG, GIFF images

31. 31 10g Geocoder  Generates latitude/longitude (points) from address  International addressing standardization  Formatted and unformatted addresses  Tolerance parameters support fuzzy matching  100% Java, open and scalable  Record-level and batch processes  Data provided by leading data vendors

32. 32 10g: Spatial Analytic Functions  Discovery based on Spatial Patterns – Explicitly materialize spatial relationships  Usage – Insurance risk analysis, crime analysis – Demographic analysis, customer profiling – Epidemiology, Facility placement – Insurance Risk analysis:  cluster house-holds based on high risk neighborhoods – Indentify business prospects across a region:  examine the average incomes across different regions of the space

33. MapViewer in Application Server 10g Intelligent Map Rendering & Analysis  Component of Oracle Application Server  Customized for the CRM/ERP Apps  Vector and Raster visualization  XML and Java Interfaces New Features:  Improved labeling and legend support  Java map client lib and JSP custom tags  Image theme & Customizable Image Renderer  PNG and Macromedia Flash mapping support  Multi-threaded geometry loading of themes  Low Memory Requirements, increased performance  New Security Features (admin login; password encryption)

34. Complying to Open Standards  OGC (GML,OpenLS  ISO TC211  W3C Consortium (XML/Web Services)  J2EE SQL/MM

35. Berik Davies Principal Consultant Oracle Spatial at UK Ordnance Survey

36. 36 Overview  Setting the scene – who are Ordnance Survey ?  Why build a new database ?  Business Requirements  Business Case  High level Architecture  Sizing and Hardware  Summary

37. 37 The early days … The Tower of London 1841

38. 38 Ordnance Survey in 2003  The National Mapping agency of Great Britain  Turnover of £ 100m (US$160m)  1800 staff  UK Government Agency with ‘ Trading Fund ’ status

39. 39 What does Ordnance Survey do ? We capture data – surveying ! We store and manage data

40. 40 … and we make products and sell them What does Ordnance Survey do ? OS MasterMap™Strategi ® Land-Line ®

41. 41 Current Environment  Existing systems lack agility  Difficult to integrate/interopera te  Very costly to maintain  Complex bespoke developments DDMSDCS Editors HSMCMS DCQTCS APS FEP DMSLMS MOSAIC GMSDef Reg GRACE CFS eDCSCSAMGDW/S Mercury (GOS) ROADS FTP RAMS GMS2 Sls & Mktng Stats Prod Sched Systems

42. 42 Why build new databases? (1) ‘Old World’ Land-Line ® ‘New World’ OS MasterMap™ (query on same area) (complete tile)

43. 43 Why build new databases? (2)

44. 44 Business Requirements  Reduce deployment costs  Reduce maintenance costs  Improve data integration  Improve systems integration  Facilitate new product opportunities  Increase business agility  Store and maintain once, use many times  Standards and COTS based

45. 45 Business Case for Oracle Spatial  An open repository for our geospatial data  Industry standard for spatial data in RDBMS  Robustness, reliability, scalability, availability  Part of the whole Oracle product family / stack  Strategic enterprise software within Ordnance Survey

46. 46 High-level overview – Maia & Mercury Maia (maintenance) Mercury (publication) Customer Data Collection Oracle 10G Spatial

47. 47

48. 48 Type of DataSize (GB) Raw Data Spatial Indexes Transient Tables required during Spatial Index Creation / Re-Build Non-Spatial Indexes History (10% change per year for 5 years) History Spatial Indexes Database Structures (system, undo, etc.) History Non-Spatial Indexes Transient Tables required during Spatial Index Creation / Re-Build Contingency at 25% Total 995 GB 175 20 40 70 350 100 10 20 200 Sizing 450 million features

49. 49 Hardware  2 x Sun Fire F6800, 12 x UltraSPARC-III 1050 MHz, 24 GB RAM  2 x Sun StorEdgeT3 disk arrays, 2 TB raw storage, RAID 5  Sun StorEdge L700 Tape system – 135 TB storage

50. 50 Summary  The Maia database, together with Oracle ® Spatial and ESRI ®’ s ArcGIS ™ will bring Ordnance Survey ’ s own internal processes into the MasterMap ™ era.  Expect financial and strategic gains from the move to Commercial Off-The-Shelf software.  The close relationship between Oracle and Ordnance Survey has brought significant benefit to both organsiations.

51. 51 Contact for further information Berik Davies Ordnance Survey Romsey Road SOUTHAMPTON United Kingdom SO16 4GU Phone: +44 (0)23 8030 5488 Fax: +44 (0)23 8079 2078 E-mail: berik.davies@ordnancesurvey.co.uk Web site: www.ordnancesurvey.co.uk

52. 52 Next Steps ….  Recommended sessions – #40340 - Leveraging Location-Based Services for Everyday Government and Emergency Management (Mon Sept 9, 4:00pm, Marriott Golden Gate A3) – Keynote: Craig R. Barrett, Chief Executive Officer, Intel (Tuesday, September. 9, 1:00 -1:45 PM) – # 40058 - Oracle Mobile and Wireless Technology Overview (Thu Sept 11 8:30 AM, Moscone Room 202) – #40340 - Leveraging Location-Based Services for Everyday Government and Emergency Management (Mon Sept 9, 4:00pm, Marriott Golden Gate A3) – # 40081 - Using Oracle Workspace Manager to Cut Costs: Case Studies (Thur. 3:15 PM, Moscone Room 131)  Oracle Spatial 10g demos – Oracle10g Location Services, Oracle DEMOgrounds

53. 53 Next Steps (cont.)  Partner Exhibition – #725 Autodesk – #1736 eSpatial – #2126 ESRI – #2408 Intergraph – # Plangraphics – # Skyline  Oracle Security Command Center Pavilion  eSpatial  PCI Geomatics  PlanGraphics  TARGUSInfo  For more information – http:// technet.oracle.com/products/spatial

54. 54 Reminder – please complete the OracleWorld online session survey Thank you.

55. A Q & Q U E S T I O N S A N S W E R S