Geospatial data management Mohamed Alkhawar IT Manager MSD, UAE Email: mkalfoot@gmail.com Eng. Mohammed Ali Kalfout Military Survey Department Delma Street (East), P O Box: 3947 Abu Dhabi, United Arab Emirates Tel: + 971 2 641 0000, Fax: + 971 2 641 5566 Email: mkalofout@uaesurvey.ae Abstract Military Survey is established in January 1974, with an initial mandate to serve military units with geospatial data and its applications. It expanded its responsibilities to serve government, private sectors and academia with accurate, standardized, and reliable geospatial data. As a result amount of geographic data generated and maintained by MSD is growing at an incredible rate. The International Data Corporation estimates data expanding approximately from 50 to 80 percent per year. The data growth rate in military survey is exceeding 100 percent annually. It constitutes 80 percent of all corporate assets. The costs associated with managing and storing ever-increasing volumes of information are also increasing dramatically. MSD is taking a lead in establishing spatial data infrastructure for the country. In order to frame the components of the SDI, Military Survey has initiated ambitious project for making National Topographic Data Base(NTDB). Because of its richness in content and accuracy, it is expected to serve varieties GIS users in the country apart its traditional users. As a result of these unprecedented growth in the volume of data, MSD has to build stable data storage and management system to ensure round-the-clock availability and reliability with efficient disaster recovery functions. This paper focuses on the steps and techniques adopted in building such an infrastructure in the MSD ensuring best performance for its users. Introduction MSD has made key investments in generating and maintaining geospatial data with a primary objective of serving the military requirements. It was serving its traditional user with geospatial data for the whole country. With this capacity MSD has emerged as a mapping agency for the UAE with the mandate of generating and maintaining national level data. MSD has initiated projects to produce and maintain up-to-date geodetic network, National Topographic Database(NTDB), DTM, aeronautical, and nautical charts for the whole country. Reliable storage and management system for this volume of data became a primary responsibility of Information Technology(IT) section. The section has framed policies for data management, and mobilized technologies for stable infrastructure providing high availability, reliability, continuity, security, and best performance for handling geospatial data sets maintained by the MSD. Old Infrastructure During the period of 1974 to 1991 Military Survey focus was on producing paper maps. In 1991 MSD shifted its focus from paper maps to digital maps. Several arrangements was prepared to ensure stable infrastructure for data storage, access, backup and security management. During these period. The network backbone was 10/100 Ethernet with switches for the same speed in each section. All MSD computers where in the same virtual Local Area Network(VLAN). MS Access Database where used in several sections. As a result viewing and editing Data was limited as well as the Database Size. Data transfer took long time and slow the network. Security also was not perfect. The available backup software and hardware where not able to handle the increased data, as a result MSD has to modify its data handling policies and upgrade its existing hardware infrastructure. Current Infrastructure: Because of the data growth, slowness in data transferring, and the implementation of geospatial data at the national level, MSD decided to improve and enhance all performance issues, to achieve stable and reliable setup as shown in Figure 1. Figure 1: MSD Network infrastructure handling Geospatial data at national level Network switches and data channels upgraded and enhanced to fiber optics backbone. MSD network is divided into several Virtual LAN’s. This will achieve high security, better performance and management. Allocating several application servers, such as Image server to handle large format raster images, license servers to control floating license for several core applications, vector database(SDE®) servers equipped with Spatial Data Engine with Oracle backend on UNIX Operating system and Windows operating system to achieve maximum data protection incase of virus attack or to block any security holes, anti-virus, etc. Upgrading all workstations to maximum performance. New policies were adopted to ensure maximum security such as router and firewall management, and in-house developed security software. . Establishing a Storage Area Network to centralize and manage geospatial data. It has been chosen to achieve high availability, scalability, reliability and increase the data handling capabilities. Implementing backup strategy over the SAN System to ensure data recovery, in the hardware and software level Versioning for multiple editing. Major migration occurred in the backend engine from MS Access to corporate Oracle® RBMS(Oracle 10g). The database is accessed and managed for geospatial application is through ESRI® ArcSDE server. The advantage of establishing the database based on Oracle is illustrated in Table 1.0. Table 1: Advantage of Oracle Backend Engine Compare MS Access Database for handling Geospatial Data Maintained by MSD Rationale Behind current implementation The MSD decided to store all geospatial database containing vector data sets in a centralized database server, so that it can be managed easily, and have it available round-the-clock. Data back-up and recovery mechanism also become more controlled as data is available in one place. To achieve the maximum performance, Storage Area Network(SAN) technology is used. Huge volume of raster data sets were separated and stored in an independent images server, which has further improved the system performance. Corporate Oracle® DBMS coupled with ESRI® ArcSDE server are used to apply more security, versioning capability, multi user editing, and unlimited number of user accesses to database. Conclusion Most components of the infrastructure is implemented. Geospatial data sets are migrated to Oracle and connected through ArcSDE. Versioning of the NTDB is created and tested successfully for multi-user editing with better performance. Clustering of server will be taking into consideration for load balancing and fail over.