A method and system for multi-level routing scheduling

Abstract

The invention discloses a method and system for multi-level routing scheduling, belonging to the technical field of network applications. The method of the present invention includes: monitoring the service, and determining the ipvs rule after monitoring the resource creation of the service; adding the identification information of the business pod of the same cluster to the integrated information to generate registration information, and registering according to the registration information; according to the request message and registration information to obtain the shortest routing address for sending business requests; when the business request and the resources in the business pod of the shortest routing address meet the business request, the resources required for the business request are delivered according to the shortest routing address of the business request. In the present invention, the appropriate routing address is determined, which improves the network access efficiency, effectively improves the service network request efficiency, reduces the network overhead caused by unreasonable routing addresses, and provides multiple services for the business system. live disaster recovery capability.

Description

technical field [0001] The invention relates to the field of network application technology, in particular to a method and system for multi-level routing scheduling. Background technique [0002] In kubernetes, service discovery and routing are realized by two resource strategies, Service and Ingress. Service mainly has three proxy modes, including: userspace proxy mode, iptables proxy mode and ipvs proxy mode; [0003] At present, the ipvs proxy mode is mainly used. In this mode, kube-proxy will monitor the KubernetesService object and Endpoint, call the netlink interface to create ipvs rules accordingly, and periodically synchronize the ipvs rules with the KubernetesService object and Endpoints objects to ensure that the ipvs status is consistent with expectations. consistent. When accessing a service, traffic will be redirected to one of the backend pods. In addition, ipvs provides more options for the load balancing algorithm, such as: rr: round robin scheduling, lc: m...

AI Technical Summary

Problems solved by technology

[0007] Under these different call modes, the capabilities of kubernetes itself can no longer be satisfied. The service routing strategy of kubernetes itself is only a simple load balancing, and it is impossible to route different service requests to specified applications in different routing methods according to the requirements of the business system. , nor can it guarantee the network efficiency of mutual access between applications, nor can it achieve multi-active disaster recovery

Images