摘要 |
A system comprises a plurality of parallel services 12 all serially connected on the output side to a consumer 14, and a method of processing messages in the system comprises: estimating the time taken for a set of messages to be processed by a service; estimating the time taken for a set of messages to be processed by the consumer; passing a first block of messages to a first service; and passing a second block of messages to a second service, where the size of the second block of messages is greater than the size of the first block of messages and will be processed by the second service in the time taken for the first block to be processed by the first service and the consumer. Staggering the amount of messages that are partitioned amongst or split between the parallel services facilitates optimised processing (Figure 3 for example). The size of the first block of messages may be determined by a minimum threshold. The estimating of durations may be continually updated in a dynamic process. |