Mobile Agent Technology

                              NETS, Inc.

                           Yong Gao, Ph.D.
                     Dundee Navin-Chandra, Sc.D.

Abstract:
        We are interested in applying mobile agent technology to the 
electronic commerce services offered at NETS, Inc. Potentially, mobile agents
can be dispatched to an electronic commerce network to perform designated 
tasks on behalf of their clients. Such a commence network must consists of a 
set of mobile agent platforms that provide secure, reliable, and controllable 
environments for mobile agent to execute, migrate and communicate.
        Our goals are two-fold. The first is to develop mobile agent-based 
applications in electronic commerce. The second is to contribute to the 
establishment of industry standards for the infrastructures mobile agents and 
mobile agent platforms.
        Mobile Agent Applications:
        At least two types of mobile agents are useful to electronic commerce:
mobile information agents and mobile transaction agents. Mobile information 
agents can be used to deliver, update or retrieve information on targeted 
server hosts. Mobile transaction agents can interact with other agents or 
humans to conduct buying or selling transactions on a network.
        Mobile Agent Attributes:
        Intrinsic properties of mobile agents include the following: (1) agent
task and goal; (2) agent identity; (3) agent credentials, and (4) agent travel
information. A standard for mobile agent infrastructure needs to specify these
attributes in a way independent of the internal implementation of agents.
        Mobile Agent Platforms:
        A mobile agent platform can be considered as a service provider to
mobile agents as well as an agent activity monitor. To support interoperability
among different mobile agent platforms, standards are needed in the following
areas: (1) a protocol for communications between an agent and the agent 
platform on which it is executed; (2) an agent transport protocol for sending 
and receiving agents; (3) a security protocol for agent activity and access 
control; and (4) an agent communication language.