OctoBot: Human Activity Orchestration System for Cybersecurity Experiment and Exercise

Cybersecurity experiment and exercise in a testbed environment required normal traffic to ensure a better result, a learning experience, and the effectiveness of the whole process. The traffic camouflages the attack traffic from the attacker by generating different types of activities and tasks that are usually generated by the human. Unfortunately, in a large-scale experiment and exercise, it is impractical to employ a large number of humans to generate normal traffic. There are several methods available to produce a single human activity from an agent, but automating it from multiple agents with multiple and different activities is very challenging. By learning and adopting the concept of BotNet, we designed an orchestration system for a human agent (i.e., bot), called OctoBot, to generate human activity that is easy to be deployed and utilized. This paper discusses primary considerations in designing and implementing our system with an example of unique bot development. It also presents deployment experience in the production testbed environment with an example use case for a virtual environment to show the potential usability of our system.