A method for dynamically distributing one or more services in a network of a plurality of computers

Method for dynamically distributing one or more services (1, 2, ..., 7) in a network of a plurality of computers (S1, S2, ..., S4), in which: a) from a past time course of the resource requirement (R i (t)) of a respective service (1, 2, ..., 7) in a predetermined time interval (Δt Ai ) a past time course of one for the respective service (1, 2, ..., required 7) resource capacity (a i (t)) is derived for the performance of a given service request by the determination of the resource capacity (a i (t)) at a time in the predetermined time interval of the time course of the resource demand ( R i (t)) in a sub-interval of the predetermined time interval which contains the time or is adjacent to this point in time, is evaluated statistically and from statistical evaluation of the resource demand (R i (t)) a resource capacity ((a t i)) determining which meets the predetermined service request in the sub-interval; b) from the past time course of the resource capacity (A i (t)) required for the respective service (1, 2, ..., 7), a chronological sequence of the respective service (1, 2, ..., 7 ) required resource capacity (A i (t)) that satisfies the predetermined service request is predicted; c) the services (1, 2,..., 7, 7) executed on the computers (S1, S2,..., S4) are distributed based on one or more optimization criteria, the optimization criterion (s) comprising the criterion that the respective computer (S1, S2, ..., S4) a sufficient amount of resources for the executed on the respective computers (S1, S2, ..., S4) services (1, 2, ..., 7) based on the predicted provide temporal profiles of the resource capacities (A i (t)) required for the respective services (1, 2, ..., 7).