摘要 |
PROBLEM TO BE SOLVED: To improve throughput related to query processings for stream data of a stream data processing system. SOLUTION: A scheduler 33 of a server apparatus 2, on the occurrence of data congestion to a query group for grouping a query 43, calculates for each query 43 constituting the query group, a load evaluation value of the query, based on at least one piece of information in input flow information and latency information of the query 43; divides the query 43 constituting the query group into a plurality of query groups so that a sum of load evaluation values of each query is substantially equal; and reallocates a plurality of query groups after the division to respective different processors 91. COPYRIGHT: (C)2010,JPO&INPIT
|