Future energy systems will present high penetrations of variable renewable energy (VRES). Due to the uncertainty that is related to the intermittency of wind and solar energy, and loads such as electric vehicles, the need for flexibility will be higher. This research identifies the possible sources of flexibility in a power system. Firstly, using statistical errors and probability theory, the required reserves of flexibility that a system needs to ensure a proper operation are quantified. Secondly, with the increase of VRES, the amount of synchronous generators is expected to decrease. This will mean that the inertial response used to keep a stable frequency will be reduced. On this paper, a method to quantify inertia and the concept of minimum required reaction time of flexibility are explained. Finally, an optimization problem is proposed to allocate the different types of flexibility taking into consideration their technical capabilities. This problem minimizes the cost of reserving different flexibility sources in a distribution grid.