| 摘要 |
A hive of computing engines, typically including clients, request handlers, process handlers, and task handlers, is used to process of a processing job in order to determine a processing result. A client assigns a processing job to a request handler, which assigns the processing job to a processing handler, which typically assigns one or more tasks of the processing job to one or more task handlers. One implementation uses one or more volunteer operations to identify the request handler, the process handler, and/or one or more task handlers. One implementation updates the request handler and/or the process handler with state information related to the partial processing of the processing job such that the completion of the processing job can be assigned to other process and/or task handlers, such as, but not limited to, in response to a detected error condition. |
