A Flow Control Method Based on Hardware Counter

Abstract

This application discloses a flow control method based on a hardware counter, which extracts the flow ID of the message; obtains the counter ID associated with the flow ID; obtains the message frame length corresponding to the flow ID; and queries the counter ID in the cache of the hardware counter ; If the query is not hit, load the message from the memory space corresponding to the counter ID to the cache; if the query is hit, go to the next step; calculate the flow state of the message in the cache according to the configuration information and the frame length of the message; according to the flow state , process the message. The flow control method provided by this application does not require an on-chip flow management module to control the flow, but uses memory as a data storage space and a hardware counter cache as a data processing space, which overcomes the relative shortage of on-chip resources and the limited data that can be stored, resulting in simultaneous management The defect of limited number of streams solves the problem of storage and processing of control data under large-scale traffic, and realizes precise control of each stream under large-scale traffic.

Description

technical field [0001] The invention relates to the field of network flow control, in particular to a flow control method based on a hardware counter. Background technique [0002] With the continuous development of modern network technology, network applications have penetrated into all aspects. The increase in network traffic is not only the increase in speed, but also the increase in the number of terminals and applications, especially the latter, which leads to the identification of network traffic control in service scenarios The number of incoming flows has increased significantly, which puts new demands on the scale of flow control. [0003] In the service scenario of network traffic control, an on-chip traffic management module is generally used to control traffic. However, due to the relative shortage of on-chip resources and the limited data that can be saved, the number of streams that can be managed at the same time is limited. It is impossible to achieve precis...

Images