MACHINE LEARNING SYSTEMS AND METHODS FOR DATA PLACEMENT IN DISTRIBUTED STORAGE

A method of determining a primary storage location for a data record in a distributed system which comprises a plurality of data stores (106a, 106b, 106c), wherein each data store is physically located in a corresponding geographic location (206, 208, 210). The method comprises initialising a machine learning mapping model using topology information of the distributed system, and determining (804) a set of training feature vectors that are derived from metadata values associated with a plurality of prior location requests. The machine learning mapping model is then trained (808) using the set of training feature vectors and a corresponding set of target primary storage locations. A location request is received (602, 702) which is associated with a data record, and which includes a plurality of metadata values, and the metadata values are processed to determine a prediction feature vector comprising a plurality of prediction feature values. The machine learning mapping model is executed using the prediction feature vector to identify (604, 704) one data store of the plurality of data stores as the primary storage location for the data record associated with the location request.