Method and system for computing resource management and billing
Patent Information
- Application Number
- TW115104514
- Authority / Receiving Office
- TW · TW
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2026-02-05
- Publication Date
- 2026-11-11
- Estimated Expiration
- 2046-02-04
Smart Images

Figure TWG2TB001921106_001 
Figure TWG2TB001921106_002 
Figure TWG2TB001921106_003
Abstract
Claims
1. A method for managing and billing computing resources, comprising: A user receives a resource request; Based on the resource request, determine whether it is necessary to fully reserve the corresponding computing power resources within a reservation period; When it is determined that reservation is required, a first billing mode is activated; when it is determined that reservation is not required, it is determined whether the requested resource has a specified reservation end time; when it is determined that no reservation end time is specified, a second billing mode is activated; when it is determined that a reservation end time is specified, based on the user's historical behavior data, an estimated cost score for each of the first, second, and third billing modes is calculated; if the estimated cost score for the second billing mode is the lowest, the second billing mode is activated; if the estimated cost score for the second billing mode is not the lowest, it is determined whether the length of the reservation period exceeds a threshold; if it is determined that it exceeds the threshold, the reservation period is divided into a first period and a second period, the third billing mode is activated in the first period, and the second billing mode is activated in the second period; and if it is determined that it does not exceed the threshold, the third billing mode is activated.
2. The method as described in claim 1, wherein calculating the estimated cost score based on the historical behavioral data further includes: The historical behavior data is input into a machine learning model to obtain a predicted execution time segment for an execution individual corresponding to the resource request; based on the predicted execution time segment, an estimated cost for each of the first, second, and third billing models is calculated; a risk cost is calculated based on the task nature of the execution individual, the probability of resource request failure, and a failure cost coefficient; based on the estimated cost and the risk cost, the estimated cost score for each of the first, second, and third billing models is calculated.
3. The method as described in claim 1, wherein the reservation period is divided into the first period and the second period, further comprising: The reservation period is divided into multiple segments, and the estimated cost score for the second billing mode and the third billing mode is calculated for each of the multiple segments. If the estimated cost score for the third billing mode is lower than the estimated cost score for the second billing mode in a segment, the segment is designated as the first period; and if the estimated cost score for the second billing mode is lower than the estimated cost score for the third billing mode in a segment, the segment is designated as the second period.
4. The method as described in request item 1, wherein, During the reservation period when the third billing mode is enabled, a minimum specification of computing power resources specified in the resource request is reserved. When the third billing mode is enabled, if the user receives a resource usage request and a full specification of computing power resources corresponding to the resource request can be provided at a point in time specified in the resource usage request, the full specification of computing power resources will be provided from that point in time.
5. The method as described in claim 4, further comprising: When the user receives an early termination request, the reserved minimum-specification computing power resources are released; a discount fee is calculated based on a remaining time during the reservation period and an overall resource utilization rate; a remaining time resource retention fee is calculated based on a usage rate for the minimum-specification computing power resources and the length of the remaining time; and the remaining time resource retention fee is subtracted from the discount fee to calculate a discounted resource retention fee.
6. The method as described in claim 4 further includes: If, after providing the full specification of computing power resources, the full specification of computing power resources are released when the user receives a request to terminate use; an actual usage fee is calculated based on the user's usage time of the full specification of computing power resources and a usage rate for the full specification; and a resource retention fee is calculated based on a usage rate for the minimum specification of computing power resources and the length of the reservation period.
7. The method as described in claim 6, wherein releasing the provided full-specification computing resources further includes releasing the computing resources of the difference between the full-specification computing resources and the minimum-specification computing resources, so as to retain the minimum-specification computing resources.
8. The method as described in request item 1, wherein, When the second billing mode is enabled, if the user receives a resource usage request and the corresponding computing resources are available at a time specified in the resource usage request, the corresponding computing resources will be provided from that time point; if the corresponding computing resources are not available at that time point, no computing resources will be provided.
9. The method as described in request item 1, wherein, During the reservation period when the first billing mode is enabled, a full specification of computing power resources corresponding to the requested resource resources are reserved.
10. A computing resource management and usage billing system, comprising: A transceiver is used to receive a resource request from a user. Storage media; The system includes a processor, coupled to the storage medium and the transceiver, and configured to perform the following: Based on the resource request, determining whether the corresponding computing resources need to be fully reserved for a reservation period; If reservation is required, activating a first billing mode; If reservation is not required, determining whether the resource request specifies a reservation end time; If no reservation end time is specified, activating a second billing mode; If a reservation end time is specified, calculating an estimated cost score for each of the first, second, and third billing modes based on the user's historical behavior data; If the estimated cost score for the second billing mode is the lowest, activating the second billing mode; If the estimated cost score for the second billing mode is not the lowest, determining whether the length of the reservation period exceeds a threshold value. When the threshold is exceeded, the reservation period is divided into a first period and a second period. The third billing mode is activated during the first period, and the second billing mode is activated during the second period. When the threshold is not exceeded, the third billing mode is activated.
Citation Information
Patent Citations
Prescriptive analytics based committed compute reservation stack for cloud computing resource scheduling
US10922141B2
Burst capacity for user-defined pools
US20140297868A1
