A method for optimizing message processing of multi-board CPUs in distributed epon communication equipment

Abstract

The invention provides a multi-board-card-CPU (Central Processing Unit) message processing optimization method in distributed EPON (Ethernet Passive Optical Network) communication equipment. The multi-board-card-CPU message processing optimization method comprises the following steps: configuring a switching chip of the distributed EPON communication equipment and rewriting a port initialization function of a switching chip driver layer; amending communication and management messages among all board card CPUs in the distributed EPON communication equipment, and adding a characteristic value to a packet header of each message; registering and filtering a message queue to the switching chip of the distributed EPON communication equipment, associating the characteristic values, transmitting the eligible messages to a corresponding message channel, and preferentially transmitting the communication and management messages among all the board card CPUs. The multi-board-card-CPU message processing optimization method realizes optimized management of the CPU messages among multiple master control cards and various service board cards aiming at the high-reliability communication demands among the multiple board cards in the distributed EPON communication equipment, thus guaranteeing high reliability of the distributed EPON communication equipment, namely the Ethernet passive optical network communication equipment.

Description

technical field [0001] The invention relates to an optimization method, in particular to a multi-board CPU message processing optimization method in a distributed EPON communication device. Background technique [0002] In the distributed EPON (Ethernet Passive Optical Network) communication equipment, the distributed management system (stacking mechanism) plays an important role to manage the status of each board, set the communication mode of the board, control the communication between the boards, collect To manage the messages of each board card, etc., it is particularly important to ensure the fast and stable communication of management messages between CPUs in a distributed system. When the flow of packets sent to the CPU is extremely large or other abnormal conditions occur on the network (network storms or various attacks), problems such as transmission congestion and loss of management packets will occur, and even the CPU has been busy processing network packets, re...

AI Technical Summary

Problems solved by technology

When the flow of packets sent to the CPU is extremely large or other abnormal conditions occur on the network (network storms or various attacks), problems such as transmission congestion and loss of management packets will occur, and even the CPU has been busy processing network packets, resulting in communication equipment failures. system breakdown

The traditional method is to limit the messages that may cause problems. On the switch chip connected to the CPU, the rate limit of various messages is set for the CPU port. For example, for IGMP, broadcast messages, unknown unicast and other special messages on the switch chip For configuration, the disadvantage is that each type of message needs to be set at the chip level at the bottom of the software, and some special message filtering switching chips may not support it.

Images