Spatial Information Grid - An Agent Framework

Spatial information grid (SIG) is a spatial information infrastructure that has the capability of providing services on-demand. In this paper, agent technology is adopted to construct a SIG framework, which contains three layers: users/applications layer, agent services layer and information layer. Different applications can get their spatial information via agent services, and agent services make the procedure of navigating and accessing spatial information transparent to users. Also, the implementation issues of the framework are discussed. 1 Spatial Information Grid (SIG) SIG (Spatial Information Grid) is a spatial information infrastructure that has the capability of providing services on-demand. SIG integrates and shares massive distributed heterogeneous spatial information resources, and provides uniform management and process. SIG is a distributed network environment, which links spatial data resources, computing resources, storage resources, software, tools and users. SIG can coordinate different spatial information resources to complete different spatial tasks and applications. In such an environment, users can present all kinds of requests for spatial data and its process, and SIG can joint distributed data, computing, network and software resources to cooperate and accomplish different users’ requests. SIG adopts a new architecture, method and technology to manage, access, analyze and integrate distributed spatial data, take full advantage of services from existing spatial information systems, so as to realize effective spatial information share and interoperation, and provides spatial information on-line analyzing processes and services. The functions that SIG should provide are: (1) The capability of processing massive spatial data. Storing, accessing and managing massive spatial data from TB to PB; efficiently analyzing and processing spatial data to produce model, information and knowledge; and providing 3D and multimedia visualization services. (2) The capability of high performance computing and processing on spatial information. Solving spatial problems with high precision, high quality, and on a large scale; and process spatial information in real time or on time, with high-speed and high efficiency. (3) The capability of sharing spatial resources. Sharing distributed heterogeneous spatial information resources and realizing interlink and interoperation at application 2 Yingwei LUO, Xiaolin WANG and Zhuoqun XU level, so as to make best use of spatial information resources, such as computing resources, storage devices, spatial data (integrating from GIS, RS and GPS), spatial applications and services, GIS platforms (such as ESRI ArcInfo, MapInfo, ... ), ... (4) The capability of integrating legacy GIS system. A SIG can not only be used to construct new advanced spatial application systems, but also integrate legacy GIS system, so as to keep extensibility and inheritance and guarantee investment of users. (5) The capability of collaboration. Large-scale spatial information applications and services always involve different departments in different geographic places, so remote and uniform services are needed. (6) The capability of supporting integration of heterogeneous systems. Large-scale spatial information systems are always synthetical applications, so SIG should provide interoperation and consistency through adopting open and applied technology standards. (7) The capability of adapting dynamic changes. Business requirements, application patterns, management strategies, and IT products always change endlessly for any departments, so SIG should be self-adaptive. 2 Agent-Based SIG Framework Agent is an autonomous, interactive, initiative and reactive computing entity in a distributed environment. Agent encapsulates some computing resources and can reach its designed goals initiatively. An agent is not only able to work on itself, but also impact environment, receive feedback information from environment and readjust its own behavior. At the same time, an agent can cooperate with other agents. Agent system reduces the restrictions of concentricity, non-openness and sequential control, provides distributed controlling, dynamic emergency processing and parallel processing services . Agent technology will provide a new thought and method for processing massive distributed heterogeneous spatial information efficiently . An agent-based SIG framework can be illustrated as figure 1. This framework contains three layers: Users/applications layer, agent services layer and information layer. Different applications can get their spatial information via agent services, and agent services make the procedure of navigating, accessing and processing spatial information transparent to users. The exchange format of spatial information between applications and agent services is XMLized (such as GML2.0). Agent services layer plays a most important role in the framework. Interface Agent presents personalized UI (user interface) for different users in different applications. Get users’ requests, communicate with the facilitator, then deal with these requests and return results to users. By the facilitator’s help, different interface agents can negotiate with each other to combine or re-plan these requests so that these requests can be replied more rationally and effectively. Facilitator administrates agents in the system, including register and message of agents. Facilitator also prompts cooperation between agents, including initializing negotiation among interface agents; acting as a bridge lying between interface agents and search agents, provider agents and process agents; scheduling search agents to Spatial Information Grid – An Agent Framework 3 navigate users’ requests; helping search agents find proper provider agent to access to spatial database, and assigning process agents to filter spatial information. Search Agent is interested in finding and locating spatial information by the users’ requests. XML repositories of spatial semantic information network that describe different heterogeneous spatial Databases support the whole searching procedure of search agents. By the facilitator’s scheduling, search agents can parallel to carry one request. As the result, search agents can acquire the constitution of spatial information and the unique locators of them. Instructed by the result of search agent, Provider Agent accesses spatial information from a right spatial database. Because the spatial databases are heterogeneous, provider agents are constructed based on different GIS product. Provider agent also translates spatial information to XMLized spatial information (such asGML2.0). Process Agent filters or queries XMLized spatial information according to the users’ requests.