Multi-cluster service diversion system
Patent Information
- Application Number
- TW115206511
- Authority / Receiving Office
- TW · TW
- Patent Type
- Utility models
- Current Assignee / Owner
- Filing Date
- 2026-07-10
- Publication Date
- 2026-10-01
- Estimated Expiration
- 2036-07-09
Smart Images

Figure TWG2TB001914536_001 
Figure TWG2TB001914536_002 
Figure TWG2TB001914536_003
Abstract
Claims
1. A multi-cluster service routing system for communicating with a user device, comprising: a plurality of computing clusters; and a service entry resolution unit communicatively connected to the computing clusters and the user device, and storing multiple cluster status information corresponding to each of the computing clusters, and multiple Internet addresses corresponding to each of the computing clusters; wherein, After receiving a service query request from the user device, the service entry parsing unit selects and transmits a target Internet address corresponding to the computing cluster providing the service from among the Internet addresses, which contains a service name for a service to be performed, to the user device. This enables the user device to generate and transmit a service request for the service to the computing cluster corresponding to the target Internet address based on the target Internet address.
2. The multi-cluster service referral system as described in claim 1, wherein, The computing clusters include a preset computing cluster and multiple non-preset computing clusters. The Internet addresses include a preset Internet address corresponding to the preset computing cluster and multiple non-preset Internet addresses respectively corresponding to the non-preset computing clusters. After receiving the service query request, the service entry resolution unit determines whether the preset computing cluster can execute the service based on the cluster status information corresponding to the preset computing cluster. When it is determined that the preset computing cluster can execute the service, the preset Internet address is used as the target Internet address and transmitted to the user device. When it is determined that the preset computing cluster cannot execute the service, the target Internet address corresponding to the computing cluster providing the service is selected from the non-preset Internet addresses and transmitted to the user device.
3. The multi-cluster service referral system as described in claim 2, wherein, The default computing cluster includes a local computing cluster, and the non-default computing clusters include multiple cloud computing clusters.
4. The multi-cluster service referral system as described in claim 1, wherein, Each computation cluster's status information includes health information, which indicates whether the corresponding computation cluster is available.
5. The multi-cluster service referral system as described in claim 1, wherein, Each cluster of operations contains a cluster status information that indicates the response delay time of the corresponding cluster of operations.
6. The multi-cluster service referral system as described in claim 1, wherein, Each computing cluster's corresponding cluster status information includes computing cost information, which indicates the resource usage cost of the corresponding computing cluster.
7. The multi-cluster service referral system as described in claim 1, wherein, Each computing cluster's corresponding cluster status information includes cluster distance information, which indicates the distance between the corresponding computing cluster and the user device.
8. The multi-cluster service referral system as described in claim 1, wherein, Each cluster of computations contains a cluster status information that indicates the number of service requests being executed by the corresponding computation cluster.
9. The multi-cluster service referral system as described in claim 1, wherein, Each cluster status information includes at least two of the following: health information indicating whether the corresponding computing cluster is available; latency information indicating the response latency of the corresponding computing cluster; computing cost information indicating the resource usage cost of the corresponding computing cluster; cluster distance information indicating the distance between the corresponding computing cluster and the user device; and load information indicating the number of service requests being executed by the corresponding computing cluster. For each computing cluster, the service entry resolution unit calculates a score based on the cluster status information corresponding to the computing cluster, and then selects the target Internet address from the stored Internet addresses based on the scores corresponding to the computing clusters.
10. The multi-cluster service referral system as described in claim 9, wherein, The service entry parsing unit can selectively choose a target allocation mechanism from a plurality of predefined allocation mechanisms. Each allocation mechanism includes a preset weight corresponding to the health information, latency information, computing cost information, cluster distance information, and load information. For each computing cluster, the service entry parsing unit adjusts the weights corresponding to the health information, latency information, computing cost information, cluster distance information, and load information according to the target allocation mechanism. The service entry parsing unit calculates the score based on the weights corresponding to the health information, latency information, computing cost information, cluster distance information, load information, the weights corresponding to the health information, latency information, computing cost information, cluster distance information, and load information for the computing cluster.
11. The multi-cluster service referral system as described in claim 1, wherein, The service entry resolution unit includes: a domain name query module for communicating with the user device and storing the Internet addresses corresponding to the computing clusters respectively; and a status information acquisition module for communicating with the domain name query module and the computing clusters, periodically obtaining the cluster status information corresponding to each computing cluster, and transmitting the cluster status information corresponding to each computing cluster to the domain name query module.
