Efficient Computing Resource Metric Method in Computing-First Network

With the continuous development of new network services and the increasing demand for computing, computing-first network (CFN) has attracted people’s attention and is gradually developing. As a method to measure the computing and storage capacity of various computing platforms, the computing resource metric plays an important role in achieving user awareness and efficient scheduling of computing resources in CFN. At present, the research on computing resource metrics is in its infancy. Most of those that only consider some static or dynamic indicators are relatively simple, which cannot guarantee the utilization of computing resources and the precision of matching computing resources. In this study, we design a hybrid metric method (HMM), which combines static and dynamic indicators to measure computing resources. This method takes the basic performance of the computing nodes and the dynamic changes in their working state into account. In addition, we also consider lots of static and dynamic indicators to enhance the comprehensiveness of HMM. The experiments and a large number of data analyses show that the metric method we propose has good improvement in the utilization of computing resources and the precision of matching computing resources.