A Pricing Method for Computing Power Services Based on Resource Tension Prediction and Electricity Price Prediction

Abstract

This invention discloses a pricing method for computing power services based on resource stress prediction and electricity price prediction, belonging to the field of cloud computing and data center operation and management technology. The method includes the following steps: Step S1. Establishing a multi-timescale computing power resource stress prediction model; Step S2. Constructing an opportunity cost mapping and comprehensive cost calculation model based on resource stress; Step S3. Constructing a differentiated pricing model and outputting the computing power service price. This method integrates long-term basic load, medium-term planned load, and short-term elastic load into a tiered occupancy ratio model, and corrects them respectively through planned execution correction coefficients, seasonal factors, special event factors, industry characteristic factors, and acceptable delay correction coefficients. Compared with existing technologies, it can more accurately reflect the actual occupancy status and dynamic changes of computing power resources. This invention can also improve the rationality and stability of the pricing mechanism, and is more conducive to the market-oriented trading and packaged delivery of computing power services.

Images