ASSIGNING MONITORING RESPONSIBILITIES IN DISTRIBUTED SYSTEMS USING OPTIMISTIC CONCURRENCY

A method performed by a service management device includes (a) receiving from a plurality of worker machines, a plurality of requests, each request from a worker machine indicating a respective work item to be assigned to that worker machine, (b) in response to receiving each request, performing an assignment attempt operation configured to (i) succeed and insert a reference to the worker machine that issued that request in a database entry associated with the respective work item if no other reference to a worker machine has yet been inserted in the database entry and (ii) otherwise report failure, and (c) returning a reply to the worker machine that issued each request indicating whether the assignment attempt operation succeeded, that worker machine configured to (i) process the respective work item in response to a successful reply and (ii) refrain from processing the respective work item in response to a failed reply.