Harmful attacks threat and counteracting them in optical networks

Optical communication is considered as safe. However, the situation looks different when attackers gain physical access to infrastructure (optical fibre cables or devices installed in the network nodes), with the purpose of eavesdropping or deterioration of such transmission (harmful attacks). Therefore, the awareness of such vulnerabilities get key meaning of design so-called intelligent networks of a new generation, which are able to protect themselves against such attacks in real time, automatically, using of ML algorithms for this purpose among others. The paper summarises different types of such attacks, togethers with outline of methods their realisation, distinguishing them on passive (eavesdropping) and active (disrupting, jamming). Protection methods based on the machine learning approach were also rediscussed at the end.