一种数据资源的调度方法及相关产品
By employing a reverse propagation algorithm and a multi-source contribution attribution model, combined with a heterogeneous data lineage topology graph, the problem of low rationality in data resource scheduling in existing technologies is solved. Dynamic scheduling of data storage nodes is achieved, improving the rationality of data resources and service stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 北京国际大数据交易有限公司
- Filing Date
- 2026-04-01
- Publication Date
- 2026-07-17
AI Technical Summary
In existing big data platforms, data resource scheduling methods based on physical attributes or access statistics are unable to perceive the logical importance of data, resulting in low rationality of data resource scheduling. In particular, small-volume or low-access-frequency but actually important data is easily misjudged as 'cold data' and evicted from the high-speed storage layer, leading to access latency and service jitter.
By introducing a reverse propagation algorithm and a multi-source contribution attribution model, positive feedback signals from terminal service nodes are obtained and backpropagated to data storage nodes in combination with a pre-defined heterogeneous data lineage topology. The global utility weight is determined based on the multi-source contribution attribution model, and resource scheduling instructions are generated, including physical storage media migration and computing task priority adjustment.
It improved the rationality of data resource scheduling, ensured the proper storage and processing of logically important data, reduced access latency and service jitter, and enhanced the rationality of data resource scheduling.
Smart Images
Figure CN121957504B_ABST