This paper discusses how to implement the procedure of information extraction from data and knowledge discovery from information in the grid computing environment

from the point of view of grid technology and information grid. Two hierarchical concepts

namely

generalized spatial information grid and special spatial information grid are proposed by the author. The former is defined as a new generation spatial information system embedded in the environment of grid technology where spatial data acquiring

updating

transferring

storing

processing

analyzing

information extracting and knowledge discovering are carried out. In order to do so

it is believed by the author that at least four functional components are needed: a smart sensor web

an intelligent data-information-knowledge processing system based on grid computing

a new generation GIS suitable to grid computing environment and Intelligent geo-service agents. The latter refers to the new generation of geographic information system in the environment of grid technology. Since conventional GISs have being developed for single person computers

four challenges caused by four types of inconsistencies are arisen when such systems are transferred to computing grid environment

they are: Challenges caused by temporal datum inconsistency

Challenges caused by geographic datum inconsistency

Challenges caused by data format inconsistency and Challenges caused by semantic inconsistency. Geospatial data representation should be re-considered if grid technology is concerned. An innovative concept

Spatial Information Multi-Grid （SIMG）

is proposed by the author and regarded as the backbone characteristics of specialized spatial information grid. The core feature of SIMG is that demarcating the studied area into different level of grids. These grids are different in size and all attribute items describing them are various and a hierarchical structure is formed where coarse grids contains several finer grids. This paper tries to put forward the definition

tasks and components of these two concepts. Some tentative approaches for solving key problems arising from them are expounded. (