CN114070663A

CN114070663A - RapidIO network based on event maintenance

Info

Publication number: CN114070663A
Application number: CN202111181049.8A
Authority: CN
Inventors: 彭文攀; 王洋; 段冰冰; 吴亦航; 张豪; 张立松
Original assignee: China Aeronautical Radio Electronics Research Institute
Current assignee: China Aeronautical Radio Electronics Research Institute
Priority date: 2021-10-11
Filing date: 2021-10-11
Publication date: 2022-02-18
Anticipated expiration: 2041-10-11
Also published as: CN114070663B

Abstract

The invention discloses an RapidIO network based on event maintenance, which comprises a plurality of modules, wherein each module is provided with more than one switching chip, a communication port of at least one switching chip on each module is connected with a processing unit, the processing unit starts a monitoring event, and the monitoring event firstly monitors the port on/off line, the port fault and the link quality of all the communication ports on the module; respectively putting the received state adding time of each communication port into the storage space of the corresponding communication port on the grid event storage; the grid event storage is provided with a storage space for each communication port of each exchange chip of the module, and each storage space is provided with two dimensional information of time and state grade. The invention monitors a plurality of communication ports on the same module by one processing unit at the same time, reduces the number of processing modules as much as possible while increasing network management, optimizes system design and reduces hardware design.

Description

RapidIO network based on event maintenance

Technical Field

The invention belongs to the field of electronic integrated network buses, and relates to an RapidIO network based on event maintenance, which is a scheme for updating and optimizing network paths in real time based on a multi-port maintenance event strategy.

Background

The avionics RapidIO network protocol is a mature generalized on-board interconnection bus which can provide reliable high-speed transmission capability and support unicast, multicast and streaming transmission modes. The RapidIO protocol has the advantages of being capable of achieving transmission capacity of more than 6G by using two pairs of receiving and transmitting pins, the speed requirement between modules is improved in a multiplied mode under the comprehensive trend of the aviation field, the low-delay requirement of the system is met by the aid of the multi-packet connection structure of interconnection, and at present, the RapidIO protocol is integrated by multiple chips.

The RapidIO bus is widely used in high-speed signal boards of aviation, aerospace, communication and the like, the communication performance of an avionic network is remarkably improved by researching the internal network technology of the bus, and the real-time routing planning inside the network becomes a new research direction at present, so that the internal network node path of the system is quickly and reliably updated, and the network routing is repaired in real time, which is an important research subject.

Disclosure of Invention

The invention aims to provide an RapidIO network based on event maintenance, which can quickly and accurately update a network topology structure through a multi-unit collaborative search scheme of a multi-port event and update and release related routing paths in real time, improve the communication capability between network nodes, reduce the probability of network packet loss and provide an efficient and quick network management scheme for a radio comprehensive technology.

The invention aims to be realized by the following technical scheme:

a RapidIO network based on event maintenance comprises a plurality of modules, wherein each module is provided with more than one switching chip, each module is provided with a communication port of at least one switching chip and is connected with a processing unit, the processing unit starts monitoring events when the last event is processed and the priority of the monitored events is highest, and the processing flow of the monitored events is as follows:

a) monitoring the port up and down lines, port faults and link quality of all communication ports on the module;

b) respectively putting the received state adding time of each communication port into the storage space of the corresponding communication port on the grid event storage; the method comprises the steps that a grid event storage is used for setting a storage space for each communication port of each exchange chip of a module where the grid event storage is located, and each storage space is provided with two dimensional information of time and state grade; the state grades are divided into port up and down, port failure and link quality.

Further, when the monitoring event finds that the state of the communication port changes, a routing collaborative search event is established, and when the processing unit finishes processing the last event and the priority of the routing collaborative search event is highest, the routing collaborative search event is started:

after finding the communication port which is positioned at the offline or the fault through the grid event searching storage, the routing collaborative searching event searches all routes leading to the communication port which is positioned at the offline or the fault from a locally generated routing table, and sends all routes to all processing units in the RapidIO network through a data packet for the processing units on other modules to close a service path leading to the communication port;

and when finding the communication port in the on-line state through the search grid event storage by the routing collaborative search event, performing routing search on the communication port in the on-line state, and sending the searched routing information to all processing units in the RapidIO network through a data packet. The routing information includes the address of the port on line, the port rate, the path to the port, and the path hop count.

Preferably, after the route collaborative search event finds a route that is not found by other units, the processing unit writes a specific number into the specific register to indicate that the route is processed, and reports the route information; after other processing units find the route and access the specific register, the path is considered to be accessed to the network and will not be processed any more.

Further, the processing module establishes a route updating event after receiving the route reported by the route collaborative search event on the other module, and starts the route updating event when the processing unit finishes processing the last event and the priority of the route updating event is highest;

and calculating the shortest route hop path and the inlet/outlet port of the shortest route by the route updating event, then generating a routing table of all target IDs in the module, and configuring the routing table on the module.

The invention has the beneficial effects that:

(1) and multi-port event monitoring, wherein a plurality of ports on the same module are simultaneously monitored by one processing unit, and event priority queuing processing is carried out, so that the number of processing modules is reduced as much as possible while network management is increased, the system design is optimized, and the hardware design is reduced.

(2) The multi-unit cooperative route searching is realized, the multiple processing units search the route in a cooperative mode, the route searching time can be shortened, the searching efficiency is improved, and the network topology structure can be calculated through the route cooperation, so that the network access or network exit events of the sub-network can be processed more accurately.

Drawings

Fig. 1 is a schematic diagram of a RapidIO network based on event maintenance.

FIG. 2 is a schematic diagram of a mechanism framework for a grid event store.

FIG. 3 is a flow chart illustrating event snooping by a processing unit.

Fig. 4 is a schematic diagram of cooperative route lookup.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples.

As shown in fig. 1, the RapidIO network based on event maintenance in this embodiment includes a plurality of modules, each module is provided with more than one switch chip, a port of at least one switch chip on each module is connected to a processing unit, the processing unit monitors states of communication ports of all switch chips on the module, such as speed, link quality, port failure, and offline on the link, by executing events, such as a residence monitoring event, a routing collaborative search event, and a routing update event, and plans a routing path according to the state of the communication port, thereby improving communication capability between network nodes and reducing a network packet loss probability. If there are more than two processing units in the module, in order to reduce the network load caused by the device service forwarding, generally only one processing unit is designated to be in the working state at the same time, and when the processing unit in the working state fails, the other processing unit is started. There can only be one processing unit at a time.

Monitoring an event:

the monitoring event is used for monitoring the communication ports of all the exchange chips on the module, and monitoring and processing the communication ports. When the processing unit finishes processing the previous event and the priority of the snoop event is highest, the snoop event is started, and the flow of the snoop event processing is shown in fig. 3:

a) and monitoring the states of the ports of all communication ports on the module, such as the on-line state, the off-line state, the port fault state, the link quality and the like.

b) And respectively putting the received state adding time of each communication port into the storage space of the corresponding communication port on the grid event storage. Each communication port of each switching chip of the module where the network event storage is located is provided with a storage space, as shown in fig. 2, each storage space is provided with two-dimensional information of time and state level. Wherein the status levels are divided into port up and down 0 (for indicating whether the communication port is in an up status or a down status), port failure 1 (for indicating whether the communication port is in a failure status), link quality 2 (for indicating the rate and link quality of the communication port), and the like.

Routing collaborative lookup event:

when the following situations exist in the RapidIO network, the routing path needs to be switched:

when a communication port fails and cannot communicate, the communication of the communication port needs to be cut off;

when the speed of a certain communication port is reduced or the link quality is reduced, reducing the communication bandwidth of the corresponding communication port;

when a single device node goes on and off a line and a single or a plurality of subnets are connected or exit from the RapidIO network through ports of a plurality of modules, a plurality of processing units must monitor the situation that the ports go on or off the line simultaneously, and a plurality of processing units are required to search a route cooperatively.

When the above conditions occur, a routing collaborative search event is established, and the routing collaborative search event is started when the processing unit finishes processing the last event and the priority of the routing collaborative search event is highest:

after finding the communication port which is positioned at the offline or the fault through the search grid event storage, the route cooperative search event searches all routes leading to the communication port which is positioned at the offline or the fault from a locally generated routing table, and sends all routes to all processing units in the RapidIO network through a data packet, so that the processing units on other modules close service paths leading to the communication port, and notify an upper layer application.

Because the process of route searching is that the processing units on the modules simultaneously search the routes of the communication ports on the line, in order to avoid repeated route searching, once the route collaborative searching event searches a route which is not searched by other units, the processing units can write specific numbers into the specific register to indicate that the route is processed, and report the route information. After other processing units find the route and access the specific register, the route is considered to be accessed and will not be processed any more after being processed, and the specific subnet access structure is as shown in fig. 4.

Routing update event:

the processing module establishes a route updating event after receiving the route reported by the route collaborative search event on other modules, and starts the route updating event when the processing unit finishes processing the last event and the priority of the route updating event is highest.

The shortest route hop path is calculated by the route updating event, the input and output port of the shortest route is calculated, then the route table of all target IDs in the module is generated, and the related exchange module is configured.

Claims

1. A RapidIO network based on event maintenance comprises a plurality of modules, wherein each module is provided with more than one switching chip, each module is provided with a communication port of at least one switching chip and is connected with a processing unit, the RapidIO network is characterized in that the processing unit starts monitoring events when the last event is processed and the priority of the monitoring events is highest, and the processing flow of the monitoring events is as follows:

2. The RapidIO network of claim 1, wherein when the monitoring event finds that the state of the communication port changes, a routing coordination search event is established, and when the processing unit finishes processing the last event and the priority of the routing coordination search event is highest, the routing coordination search event is started:

when finding the communication port in the on-line state through the search grid event storage by the routing collaborative search event, carrying out routing search on the communication port in the on-line state, and sending the searched routing information to all processing units in the RapidIO network through a data packet; the routing information includes the address of the port on line, the port rate, the path to the port, and the path hop count.

3. The RapidIO network based on event maintenance as claimed in claim 2, wherein after the route collaborative search event finds a route that is not found by other units, the processing unit writes a specific number into a specific register to indicate that the route has been processed, and reports the route information; after other processing units find the route and access the specific register, the path is considered to be accessed to the network and will not be processed any more.

4. The RapidIO network maintained based on events of claim 2 wherein the processing module establishes the route update event after receiving the route reported by the route collaborative search event on the other module, and starts the route update event when the processing unit finishes processing the last event and the priority of the route update event is highest;