Abstract: A system and method are provided for operating a pool of workers within a distributed computing system, wherein each of multiple computers host one or more workers. Each computer also hosts ledgers for one or more originators of tasks that are submitted to the pool for processing. A ledger identifies tasks from the corresponding originator that have been submitted to the pool and that have not yet been completed. When an originator submits a new task, the system (e.g., a coordinator) identifies a maximum number of concurrent tasks permitted the originator and determines, with reference to the originator's ledger, whether the task can be submitted to the pool without exceeding the limit. If so, the task is assigned to an idle worker (if any) and is added to the originator's ledger. After completion, the task is removed from the ledger. Obsolete or stale entries are periodically purged from originators' ledgers.

Abstract: A method and system for managing ad campaign data. According to one embodiment, an application receives a request to store data associated with an ad campaign, and responsive to the request, stores one or more output files and context information associated with the data, the context information assembled according to a pre-defined data format defining a role of the data within the ad campaign.

Abstract: A method and system for managing ad campaign data. According to one embodiment, an application receives a request to store data associated with an ad campaign, and responsive to the request, stores one or more output files and context information associated with the data, the context information assembled according to a pre-defined data format defining a role of the data within the ad campaign.