31-07-2012, 11:33 AM
A Web Agent for Automating E-Commerce Operations
A Web Agent for Automating E-Commerce Operations.pdf (Size: 243.34 KB / Downloads: 144)
Abstract
The great amount of information that a user handles in
their typical transactions on the Internet –identifiers,
addresses, credit card numbers, among others-, makes it
necessary to have a mechanism which facilitates the
secure management of that information, and its use in a
convenient way on a daily basis. The main goal is to
favour the B2C –Business to Consumer- e-commerce
activity by creating a friendly and secure environment.
This paper introduces an Internet navigation assistant,
implemented as a generic, extensible and dynamicallyupdatable
personal agent. This agent is included into a
distributed architecture which is composed by a set of
secure web servers and on the client side, one agent per
user machine, developed as a Microsoft Internet Explorer
toolbar. The agent includes, among others services, an
automatic-filler which learns from users’ activity, and
secure payments services, like a virtual-card generator or
an Europay’s SPA/UCAF wallet.
Introducción
The great amount of information that a user handles in
a typical Internet transaction usually converts the
experience of B2C purchasing into a tedious, messy and
error-prone process. The typical information that a user
must deal with includes the following:
– Personal information for registration in web sites:
name, surname, home and email address, etc.
– User identifiers and passwords to access
authentication-enabled sites.
– Information related to making orders, like credit card
numbers, address for invoice, delivery addresses, etc.
An application which facilitates the automatic
management of such information will indeed be of great
help for the user in the purchasing process, translating it
into a much friendlier experience.
System architecture
To manage the necessary information of the users, the
system makes use of a distributed architecture, based on a
group of secure web servers. This offers two main
advantages: first, it allows for a complete user mobility,
and secondly, it minimizes the exposure of sensible data,
thus avoiding several security risks associated to elements
such as local file storage and taking into account others
like browser cached information.
Agent architecture
The agent installed must have access to the objects
which have been loaded into the user browser –pages,
forms, form fields, and so on-, monitoring certain events
produced on them, accessing and modifying some of their
properties. Therefore the agent must be coupled to the
browser, suggesting it should have a different
implementation depending on the particular browser to be
considered. In this first implementation we have chosen
MSIE, due to its broad spread among the potential users
of the system presented here. The type of plug-in selected
has been a toolbar, which accomplishes both technical
and interface requirements.