System comprising execution nodes for executing schedules

Abstract

In a system (1) with storing means (2) for storing schedules (200, 210, 220, 230) and comprising nodes (3-6) for executing the schedules (200, 210, 220, 230), first and second nodes (3, 4) each comprise checking means (302, 402) for checking the storing means (2) in dependence of a node's load and collecting means (303, 403) for collecting schedules on the node's own initiative. This all is done without the nodes (3, 4) being monitored by a monitoring process of a system supervisor. At least one so-called single point of failure in the form of a monitoring process has been removed. Marking means (304, 404) mark the schedules in the storing means (2). The nodes (3, 4) comprise event driven scheduler (300, 400) for controlling the means (302-304, 402-404). The checking means (302) of the first node check the storing means (2) a first predefined time-interval ahead of an actual schedule and the checking means (402) of the second node check the storing means (2) a second predefined time-interval ahead of an actual schedule, the first predefined time-interval depending on the first executing node's load and the second predefined time-interval depending on the second executing node's load, to execute the schedules in a load balanced way.

Description

Technical field [0001] The present invention relates to a system including a storage device for storing a schedule and an execution node for executing the schedule. [0002] The node may be, for example, a server node forming a part of a server system or a network node forming a part of a network system, for example. Background technique [0003] The known prior art system comes from US 2003 / 01 / 0172212 A1, which discloses a cycle judgment system. As disclosed in paragraph 0005 of US 2003 / 01 / 0172212 A1, the cyclic judgment system is defined as continuously and cyclically allocating resources to many requesters. [0004] The known system is flawed, especially because of the fact that the cycle judgment system needs to monitor the process to monitor the nodes. Such monitoring process makes the system relatively complicated. In addition, the monitoring process forms a so-called single point of failure. In the event that the monitoring process fails, the entire system fails. Summary ...

AI Technical Summary

Problems solved by technology

Such a monitoring process makes the system relatively complex

In addition, the monitoring process forms a so-called single point of failure

In the event of a monitoring process failure, the entire system fails

Images