Depending on existing availability requirements, appropriate measures must be taken to avoid failure of partial or complete systems. Due to the great number of systems to be protected, there exist an at least equally great number of technical and organisational measures to minimize the probability of a system failure. A frequently used method consists in the redundant design of critical systems. Servers, for instance, have two power supply units, i.e. in the event of a failure of one power supply unit, the server can still be supplied with power through the remaining unit. Very critical systems up to complete data processing centres frequently have a double-redundant design.