IDSOJ: An Intrusion Detection System in Online Judges for evaluating programming exercises

IDSOJ is an intrusion detection system in the environment of a self-correcting online judge for programming student practices. This system detects the entry point of the different attacks that the students can carry out on the judge, informing the user with generated explanations about the mentioned attacks, as well as indicating several ways to combat or avoid them in the future. These attacks oscillate between access to the solutions of the practices of their colleagues, the possibility of changing the verdict of the judge or the date of submission (to approve suspended or not presented practices) and identity theft of a user with teacher role, among others. This paper presents the characteristics of the IDSOJ and shows the results obtained with it.