Method and equipment for determining key node

Abstract

The invention discloses a method and equipment for determining a key node. The method comprises the following steps of: obtaining operation times corresponding to various nodes in a system; determining a link having the longest operation time from an initial node to a target node; counting optimizable durations corresponding to various nodes included in the link; and selecting a node, the optimizable duration of which accords with a pre-set condition in the link, as the key node. By means of the method and the equipment disclosed by the invention, the optimization work is guided by searching the key node causing delay of the target node; therefore, the optimization efficiency is increased; and the task delay phenomenon is relieved.

Description

technical field [0001] The embodiment of the present application relates to the field of communication technology, and in particular to a method and device for determining a key node. Background technique [0002] A large-scale task scheduling system often contains tens of thousands of working nodes, and the nodes depend on each other to form a huge task tree. When the system is running, it needs to judge from the root node layer by layer which parent tasks of the tasks are all completed, and the tasks that meet the conditions Start running. In the task scheduling system in the data warehouse, tasks are often scheduled and executed on a daily basis, but when the dependencies of nodes become increasingly complex and the absolute number is also increasing, some downstream tasks have been delayed to a very serious state, even It will become the task calculation on the third day, and the task of the second day will be blocked at the same time. [0003] A key node refers to a c...

AI Technical Summary

Problems solved by technology

In the task scheduling system in the data warehouse, tasks are often scheduled and executed on a daily basis, but when the dependencies of nodes become increasingly complex and the absolute number is also increasing, some downstream tasks have been delayed to a very serious state, even It will become the task calculation on the third day, and the task of the second day will be blocked at the same time

[0003] A key node refers to a certain node in the task scheduling system. Optimizing this node can significantly improve the performance of the scheduling system. In the prior art, the optimization of the task scheduling system is usually to directly optimize the node with the longest running time among all nodes. , but due to the intricate dependencies in the large-scale task scheduling system, the node with the longest running time among all nodes is not necessarily related to the task with severe delay, and the node with the longest running time is not necessarily the node with the highest degree of optimization, so only Optimizing the node with the longest running time among all nodes lacks pertinence and cannot guarantee the efficiency of the optimization work, so it cannot guarantee the improvement of the efficiency of the entire task scheduling system, and the low efficiency of the scheduling system will lead to continuous delays in tasks, and eventually Severe data lag issues

Images