Abstract: A method for distributing non-transactional workload in the form of online analytical processing (OLAP) load across multiple database servers comprises receiving, at a scheduler server, a plurality of asynchronous work requests, adding, by the scheduler server, each of the received asynchronous work requests to a work list queue as a work item, and determining, by a scheduler module of the scheduler server, a first work item from the work list queue to select for performance by the first thread of the application server. In accordance with one or more preferred implementations, the first work item is selected utilizing a methodology which determines a number of work items in the work list queue having a processing status that are assigned to a database server, and based thereon determines whether an additional work request can be assigned to that database server without exceeding a request threshold for that database server.