Review of the Main Security Problems with Multi-Agent Systems used in E-commerce Applications

Abstract

The ability to connect to the Internet from a wide variety of devices such as smart phones, IoT devices and desktops at anytime and anywhere, produces a large number of e-commerce transactions, such as purchases of clothes, ticket entrances for performances, or banking operations. The increasing number of these transactions has also created an increase in the number of threats and attacks by third parties to access user data banks. It is important to control the access procedure to user data so that the number of threats does not continue to grow. To do so, it is necessary to prevent unauthorized access, theft and fraud in electronic commerce, which is required to ensure the safety of these transactions. Many e-commerce platforms are developed through multi-agent-systems because they include certain advantages to control the product, resource management, task distribution, etc. However, there are a number of threats that can jeopardize the safety of the agents that make up the system. These issues must be taken into account in the development of these multi-agent systems. However, existing methods of development do not cover in depth the issue of security. It is necessary to present and classify the potential security flaws of multi-agent systems. Therefore, the present research presents a review of the main vulnerabilities that occur in multi-agent systems responsible for managing e-commerce applications, as well as the proposed solutions to the major security problems on these platform systems. The main conclusions provided by this research is the need to optimize security measures and enhance the different security solutions applied in e-commerce applications in order to prevent identity theft, access to private data, access control, etc. It is therefore essential to continue to develop the security methods employed in applications such as e-commerce as different types of attacks and threats continue to evolve.