and 
bility 
Level 
°d by 
ween 
'd on 
built 
DSS. 
ction 
s of 
d by 
; the 
ision 
- for 
je of 
se, is 
oned 
ently 
asic 
hies. 
blem 
er is 
  
provided in order to resolve the heterogeneity between the 
different SDSS databases. In this section we only emphasis on 
the system architecture for resolving the heterogeneity. In 
section 4, aspects of heterogeneity presented. 
LOCAL SERVER (LS): 
The local server represents the gate for communication of the 
local database with the global server. The local server contains 
data that each participating basic database is willing to share 
with the federation. The sharable data is supported by a 
metadata which contains information about the data stored in the 
database [Bishr Y., 1996]. 
The metadata specifies the characteristics of the data items: 
format, quality, collection and processing procedure, etc. The 
metadata should also contain schema and mapping definitions to 
enable the abstraction of the sharable data to a global export 
schema. 
The role of the LS is to put at the disposal of clients a common 
object-oriented interface on top of the local DBMS. The view is 
then object oriented and is implemented through methods. In a 
sense, the LSs are abstractions of the local DB. It is responsible 
for accessing and retrieving data as requested by the Client 
through the Global Server. For the present work, the server 
hosts elementary databases of single data type, e.g. soil, 
hydrology, land cover/use, relief, etc., at a detailed map scale. 
GLOBAL SERVER (GS) 
The global server has a global external schema which provides a 
mechanism for resolving the different aspects of heterogeneity, 
section 4. In the proposed architecture, each level of decision 
making: i, j, or k, from now on called client, is provided with a 
GS, i.e., GS; or GS; or GS,, which support all users at their 
underlying level. 
To support each level, the corresponding GS links its client with 
the LSs. Upon the receipt and acceptance of the client's request, 
the following operations are performed: 
* Analyze the request to identify and locate the required data 
items and send the corresponding messages to the 
appropriate LSs. 
* Receive the data sets from the LSs and process them to 
provide the adequate data items. 
* Reply to the client by sending the requested data. 
Additionally, the global server has the following tasks: 
* Control the transactions with the clients and data sources. 
* Maintain the global directory, i.e., information about the data 
available within the federation: location, information on 
Specific data sets, ownership, format, cost, etc. This is 
achieved by storing a comprehensive metadata in the global 
server. 
* Execute data conversion: units, formats, etc. 
MULTILEVEL SERVER 
The multi-level server is responsible for linking the different 
decision making hierarchies to establish the corresponding 
feedback between the three decision levels in terms of data, 
knowledge, and decisions necessary for their activities. Its tasks 
667 
are: 
e Control the communication between clients (management 
levels), 
e Access and retrieval from the corresponding DSS database in 
a specific level. 
In this section, the system architecture and the functionality of 
each component are explained. In the next section the 
components of the DSS are outlined. In section 4, more 
emphasis is given on the different aspects of heterogeneity of the 
databases which support such an architecture. section O shows 
methodology followed for implementing such an architecture. 
Then the paper is concluded in section 6. 
3. ARCHITECTURE OF THE SDSS 
The proposed system, Figure 3, is a hybrid system which 
incorporates an expert system, a GIS software, a remote sensing 
software, and application Models. The selection of components 
of the proposed system is the result of the classification and 
definition of the problem types to be solved in WSM. The 
system integrates the supporting databases of each of those 
components into a single data model. The system has the 
following Components: 
  
   
  
  
Applic. 
  
  
  
Inference Engine 
Rule Base 
  
  
  
Object Schema 
  
  
  
Global Server 
  
     
  
(one level) 
Global Server 
   
  
(multi-level) 
   
Figure 3 Components of SDSS 
GIS Component 
GIS functionality is accessed through its interface with the DSS. 
The possibility to use geographic information systems enables 
DSS users to handle spatial and non-spatial data from different 
disciplines for the purpose of spatial data management, analysis 
and presentation. 
RS Component 
Remote sensing capabilities are accessed through its interface 
with DSS. This capability provides the benefit of capturing and 
analyzing remote sensing data and the opportunity to perform 
monitoring operations of particular aspects of the environment 
relevant to watershed management. 
Application software 
In order to provide specialized analysis in watersheds, there are 
several simulation models which are designed to analyze 
different aspect of watersheds, e.g., soil erosion and water 
quality. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996