In recent years, numerous attacks have appeared that aim to steal secret information from their victim using the power side-channel vector, yet without direct physical access. These attacks are called Remote Power Attacks or Remote Power Analysis, utilizing resources that are natively present inside the victim environment. However, there is no unified definition about the limitations that a power attack requires to be defined as remote. This paper aims to propose a unified definition and concrete threat models to clearly differentiate remote power attacks from non-remote ones. Additionally, we collect the main remote power attacks performed so far from the literature, and the principal proposed countermeasures to avoid them. The search of such countermeasures denoted a clear gap in preventing remote power attacks at the technical level. Thus, the academic community must face an important challenge to avoid this emerging threat, given the clear room for improvement that should be addressed in terms of defense and security of devices that work with private information.