Int. J. Advanced Networking and Applications Volume: 13 Issue: 04 Pages: 5024-5028(2022) ISSN: 0975-0290 5024 Protecting Mobile Agent against Man-In-The- Middle Attack: The Dummy Agent Model Mohammed Suliman Faculty of Computer Studies, Arab Open University, Saudi Arabia. Email: msuliman@arabou.edu.sa Bandar Alluhaybi Faculty of Computer Studies, Arab Open University, Saudi Arabia. Email: b.alluhaybi@arabou.edu.sa ------------------------------------------------------------------- ABSTRACT--------------------------------------------------------------- With the fast growth of the Internet in the world, playing an important part in our daily activities and with different uses being personal or in business. The idea of the internet is based on transferring data between computers. Agent-Based Software Technology (ABST) is one of the most popular technique to exchange messages and perform the task between devices over the network. The Software Agent (SA) become vulnerable to attacks by Man-in-the-Middle (MitM) during the journey in the network. To secure the software agents against MitM attacks we propose a Dummy Agent Model (DAM) by confusing the attacker and making it difficult to differentiate the real agent with dummy agents. Dummy agents selection model is used to select dummy agents from the historical agent database for every real agent making it more difficult to attacker to identify the real agent among dummy agents. The Dummy Agent Model is proposed as a better approach to active attacks such as DoS, Collision and Alternation attacks. Keywords – Attack, Destination Machine, Man in the Middle, Software Agent, Source Machine. -------------------------------------------------------------------------------------------------------------------------------------------------- Date of Submission: Nov 07, 2021 Date of Acceptance: Dec 14, 2021 -------------------------------------------------------------------------------------------------------------------------------------------------- I. INTRODUCTION 1.1 Software Agent and its work. A software Agent (SA) is an autonomous process that behaves like a network user program on a computer network [1]. Agent-Based Software technology (ABST) is found to be one of the significant technology used to manage and execute jobs over the network (internet). In contrast with other technologies such as remote procedure call (RPC)[2], Message Passing (MP)[3], ABST found to be better in terms of executing user programs for a high number of users and high size data, with less access latency and tuning time levels and better throughput in execution in distributed systems[1], [4]. Multiagent work more similar to multithreaded technology in the network [5]. One of the issues in the network when transmitting big data especially when it is multimedia data (audio, images and videos) can be addressed by using agents, where the SA can be shifted from Source Machine(SM) to Destination Machine (DM) to perform tasks in DM. The SA process the big data in DM and send only the results to SM as required, which significantly reduces the network overhead [6]. This is also illustrated in Figure 1. Figure 1. ABST view Further, a software agent carries four different type of information with it, which is travel information that includes SM and DM details, code of the agent, data used by agent code, state of agent which includes results executed task and other task execution related information. Figure 2 describe the SA parts. Security of information passed between devices is always a concern in e-learning using IoT [7]. Figure 2. Software agent parts 1.2 Problem Statement Software Agents are autonomous and are created, located in the network. These agents can move freely from machine to machine in a network. For such agents, the most important features to hold are Mobility, that agents can move from one SM to DM, doing all their job and return to home(SM); agent huskiness and portability that agents can work in a different environment with different resources and data; and agent transparency and accountability to the SM giving all the process details and results to the agent owner machine (SM) [6], [8], [9]. Software Agents are powerful with these features, but it is also vulnerable to attack in the network during their journey, to the DM, by the DM.