2004 
  
lunay 
lided 
patial 
4 
S: A 
ithm. 
tural 
ling, 
igust 
rated 
10! 
phic 
rine 
Tc 
'oint 
onal 
) 
>. 
992. 
ITC 
sitüt 
ata 
ing, 
)our 
iate 
may 
The 
and 
sep: 
-nal 
AGENT TECHNOLOGY AS A SOLUTION FOR NETWORK-ENABLED GIS 
Saeid M. Kalantari * , Ali A. Alesheikh ^ 
“ Graduate student of master, Dept. of GIS Eng. sm_kalantary@yahoo.com 
8 Assistant Professor, Dept. of GIS Eng. alesheikh@kntu.ac.ir 
° Faculty of Geodesy and Geomatics Eng., K.N. Toosi University of Technology, Vali asr St, Tehran, Iran, P.C. 1996715433 
Tel: +98 21 8789357, Fax: +98 21 877 9476 
KEY WORDS: GIS, Mobile, Object, Distributed, Networks, Artificial Intelligence 
ABSTRACT: 
With widespread availability of World Wide Web (www) and the acceptance of the intranet in various organizations, a new generation of 
GIS is burgeoning as Distributed GIS (DGIS). DGIS has open architecture distributed computing and high level mobility in mobile 
application and services. The Intranet involves in integrating a number of functions that are distributed physically or logically. Such 
distribution makes many challenges like inconsistent information, multiple information sources, decentralized control, conditional 
operation with failure and many other problems. 
To interact with such systems and tackling mentioned problems in DGIS, new paradigm of software engineering and artificial intelligence 
combination has been introduced as agent technology. This paper introduces a new solution for distributed GIS using mobile agent 
technology. The concepts have been implemented in a case study and the results of test are evaluated scientifically. 
1. INTRODUCTION 
The network-based GIS introduced a new paradigm that 
enabled a broader variety of GIS applications and expanded 
GIS users. The new and innovative architecture delivered GIS 
for wide distribution across enterprises and organizations and 
made GIS accessible over the ubiquitous network. 
1.1 Network-based GIS architectures 
The development of network technology facilitated the 
applications of network-based GIS in two network levels: the 
Enterprise level and the Internet level. An enterprise network 
might be either the Local Area Network (LAN) or the Wide 
Area Network (WAN) and the Internet is also referred to as the 
World Wide Web (WWW). More specifically, GIS software 
exists in the following forms of network: 
1.1.1 Host-Terminal network: This is the old model of 
networking in which a mainframe computer acts as the host and 
many terminals are used to access the data and GIS functions. 
Since every computation is calculated in the host and the 
terminals are only used for display and interactions, this model 
has very high performance requirements to the host. The major 
problem of this model is its slow response speed, high cost, and 
difficulty of development. Current networked GIS discarded 
this model (S. M. Kalantari, 2004). 
1.1.2 Client-Server network: This model of network is widely 
exist within enterprises, in which some computers act as 
servers as well as others act as clients. The server computers 
usually have more power than the client and manage the 
centralized resources (figure 1). Different from the old host- 
terminal model, the client machines in this model also have 
some resources and computational power that might be used to 
709 
relieve the load of servers. This characteristic of client-server 
network made it faster, more flexible and less costly than the 
host-terminal network. Actually, the client/server network is 
the major form of network in the enterprises (Yuan S.. 2000). 
  
Clients 
     
  
Client/Server 
Figure 1: Client/Server architecture 
1.1.3 Distributed architecture: In the third (Distributed) 
architecture. (figure 2), GlServices are built upon a more 
advanced networking scheme. The significant difference is the 
adoption of distributed component technology, which can 
interact with heterogeneous systems without the constraints of 
traditional client/server relationships (McCarty B,1999). Under 
a distributed architecture, there is no difference between a 
client and a server. Every GIS node can act as a client or a 
server based on the task. A client is simply defined as the 
requester of a service. A server (likewise) is simply the 
machine that provides the service. This architecture permits 
dynamic linkages between data and software. In fact. the 
architecture is very similar to what 1s called “peer-to-peer”