Flow scheduling method and related device

Abstract

The invention discloses a traffic scheduling method, a VNF cluster shrinking and expanding method and related devices, belonging to the field of communication technology, wherein a traffic scheduling method includes: a network controller acquires the quantity of all data packets entering a switch, The table obtains the data of the first flow entry; the VNF manager obtains the status of each VNF instance; the network controller and the VNF manager summarize the collected information into the orchestrator, and the orchestrator calculates the best Large flow scheduling strategy: the network controller sends the second flow entry to the large flow flow table, and when the second group of data packets enters the switch, the second group of data packets is matched with the second flow entry ; This method starts from the flow size distribution in the flow, and proposes to use the flow table to accurately schedule large flows. This method can achieve load balancing of instances within the cluster when dealing with massive flows in large-scale scenarios, and at the same time reduce flow migration and The overhead caused by instance scaling.

Description

technical field [0001] The present invention relates to the field of communication technologies, and in particular, to a traffic scheduling method and related devices. Background technique [0002] In carrier or enterprise networks, these various types of network functions are usually deployed, such as firewalls, load balancers, network address translators, etc. These devices can help enhance network functions and security. However, because network function devices with different functions are usually dedicated hardware, the differences between different hardware devices make enterprises or operators need a lot of overhead and investment in the process of deployment, update, and maintenance. Facing the rising maintenance and operation cost, NFV (Network Function Virtualization, Network Function Virtualization) technology emerged as the times require. By decoupling network functions from hardware devices, all network functions are realized through software, so that they can r...

AI Technical Summary

Problems solved by technology

However, this method is prone to overflow of the flow table, resulting in performance degradation

Secondly, OFM uses integer linear programming to solve the optimal flow scheduling and scaling strategy. When the number of instances in the VNF cluster is large and the number of flows is large, the number of variables in the integer programming model will also increase exponentially, resulting in integer Planning model solve time impossible to accept

In summary, OFM cannot be applied in large-scale scenarios

Images