|
Promoting Distributed GIServices Using Mobile Agent Technology
2. Distributed GIS
Distributed GIS is a subset of distributed system so whole coordination and requirements of it must be considered. In addition because of special data type and processing tools in this kind of distributed system, other GIS specifications are necessary. The ideal model of distributed GIS is “Geodata anywhere and Geoprocessing anywhere” [2]. Geodata and Geoprocessing do not have to be in the same side of the network. Four important part of conceptual architecture of distributed GIS are client, network, geodata and geoprocessing .Network is located in the centre of this architecture and other parts connect via network(see figure 1).  Figure 1: Geodat and geoprocessing distribution in the network [2]. Almost all of current distributed GIS products are still only for geodata publishing purpose .Geoprocessing tool keeps as a week point [2]. Although there are some distributed GIS focused on geoprocessing tools but all of them challenge with considerable problem .Most of them implement geoprocessing tools near the data. Assuming the possibility of the geodata and geoprocessing tools in one locationbe presented by P, no matter it is at a client site or at a server site. If D represent geodata, and F indicate geoprocessing tools (Functions), we have D: geodata, F: geoprocessing null: none of geodata or geoprocessing tools. Then, we have: D = {D, null} and F = {F, null} So P = D × F = {(D, F), (F, null), (D, null), (null, null)} In which: (null, null): Only exist at the client side, which means that neither geodata nor geoprocessing tools exist at the client side. The client in the typical client/server GIS model fit this case. In this case, the client is a thin client. (D, F): Typically exist at the server side in a typical client/server GIS architecture, which means that all geodata and geoprocessing functions are provided by the server. Since the server has everything, its clients should be thin clients (see the (null, null) case). Standalone GIS systems fit this case. (F, null): The site has all geoprocessing components but none of geodata. The geodata are separated from the geoprocessing and distributed in other places. (D, null): This may probably exist as a data center. Geoprocessing may be provided by other sites. Some products of geodata and geoprocessing are located in a same side in server (thin client). So when a client sends a request to server because of high CPU usage of geoprocessing tools and also high memeory load of geodata (for example shortest path) other clients must be waiting until server finish the process. This causes high traffic in the network. In another was if geodata sent to geoprocessing side high volume of data must be uploaded in the network and finally we have high traffic in such networks. This kind of connection can be seen in applets. Although after downloading the client can interact with system and data easily and quickly but network traffic during the period of downloading is failed. There is another manner in this kind of systems which sending geoprocessing tools to geodata side .But in this manner a geoprocessing tool must be ready and move it in a manner of run software through network to geodata side. This can be done by new paradigm of computer science entitled as agent technology. 3. Agent Technology Agent concept began by an idea of non-human agencies. It's is realized by constructing robots. Nowadays it is implemented as soft robot, living and doing its task within computer world. "An agent is a computer system suited in some environment and that is capable of autonomous action in this environment in order to meets its design objectives"[3]. There are several characteristics for an agent ,some of them are ideal characteristics and are far from reality. But some characteristics like mobility, communication ability, reactivity and inferential capability can enhance GIS applications in various fields. |