KR100462474B1 - Message exchange device and method for IPC(inter processor communication) using the packet ring - Google Patents

Abstract

The present invention is to configure the Inter Processor Communication (IPC) configuration in the form of Packet Ring in a distributed system such as a large capacity router system or a switch system having a 10 Gigabit Ethernet port, for example. The present invention relates to an IPC message exchange apparatus and method using a packeting configuration that enables IPC transmission without an Ethernet switch.

To this end, the present invention can be applied to a scalable router by configuring an IPC configuration in the form of packeting for reliable system operation, changing an IPC path when a failure occurs in the system, and providing a routing function and a packet processing function. Allows development of a decentralized, decentralized system. By using the present invention, it is possible to increase the stability of the system and to simplify the system configuration further, thereby reducing the overall cost.

Description

Message exchange device and method for interprocessor communication (IPC) using the packet ring}

The present invention is, for example, in a high-capacity router system having a 10 Gigabit Ethernet port, an apparatus for inter-processor communication (IPC) as well as communication between the internal line card module and the central controller as well as other router systems. And to a method.

That is, the present invention is to configure the Inter Processor Communication (IPC) configuration in the form of Packet Ring in a distributed system such as a router system or a switch system to enable IPC transmission without an Ethernet switch. An apparatus and method for exchanging an IPC message using a packeting configuration are provided.

The recent increase in demand for Internet services has accelerated the surge in Internet traffic due to the demand for stream-oriented application services such as VoIP and multimedia services and the demand for mobile Internet services. This increase in services has increased the use of high-speed switch equipment and high-speed router equipment, and research has been continuously conducted to efficiently develop such equipment.

Looking at the configuration of a conventional network equipment with reference to Figure 1, in order to escape the performance limitations of the network equipment using a conventional general microprocessor, it is possible to implement high-speed packet switching and routing, Layer 3 switching and deep sub The use of a network processor (NP) 110a .. 110d, which is a network-only processor capable of supporting Diffserv. As hardware, has become common.

Hereinafter, in the present specification, the same member number is used for constituent members having the same or similar function, and a lower case alphabet is added after the member number to distinguish corresponding constituent members. Therefore, in the case where only the member number is used without the lowercase alphabet, all the corresponding constituent members shown in the drawings will be indicated.

The network processor 110 generally operates like a host processor (CP) 109, which is a general microprocessor, in order to separate a function for packet processing from a control and management function. As an interface 117 for this purpose, a PCI bus or the like is provided. While the network processor 110 functions to support hard-wired speeds such as header parsing, pattern matching, bit filter manipulation, table lookup, packet modification, and traffic management, the CP 109 initializes the system. Performs management and control functions such as adding entries to the forwarding table, managing multicast groups, changing thresholds, etc. in buffer management.

Recently, IP forwarding, protocol conversion, quality of service (QoS), security, traffic bandwidth allocation, and VoIP functions are added as additional functions of the network processor 110. To this end, upper layer (L4 to L7) Requires packet processing For this purpose, the CP 109 also supports external upper protocol stacks L4 to L7.

In the case of a general router using the network processor 110, a main processor (MP) 101 dedicated to processing routing related protocols and maintaining a routing table, a switch fabric 113 for switching input / output packets, a network interface, It consists of line card modules 106 that handle the forwarding function.

In this conventional general router structure, a CP 109 is provided for each line card 106 so that the CP 109 controls the network processor 110 in the line card 106 and the n host processors 109. MP 101 again manages through the IPC (Inter Processor Communication) function using the Ethernet switch 105 in the system. This is because a large-capacity router system composed of a plurality of line card modules 106 requires inter-processor communication (IPC) for transmitting and receiving necessary information with a processor of another module.

In the structure of FIG. 1, which is a conventional general structure, high-speed packet forwarding is performed by a network processor of a line card, and a host processor is responsible for routing calculation and table updating, so that content having routing information is communicated through an IPC pass 111. This is done and packet processing is being developed as a distributed router system using a path 112 towards the switch fabric 113.

One of the most important operations in the distributed router system architecture is the management of the routing table, along with the calculation of the route shortest path in the main processor, packet classification and scheduling in the network processor, efficient destination search, and updating of the forwarding table for QoS function support. .

A reliable separation of routing and forwarding functions is very important to make this operation reliable. However, in order to solve problems such as frequent IPC traffic increase between the network processor and the main processor when a route flap occurs due to instability of the network, and to solve the obstacles quickly, duplication of routing processing is required so that two main processors (101a, 101b) is present. That is, when one main processor 101a fails, the other main processor 101b replaces the routing function. These main processors also mate with Ethernet switch 105 and Ethernet connections 103a and 103b for IPC.

In FIG. 1, reference numeral 107 denotes a physical layer connector PHY, and 108 denotes a transformer connected to a connection port.

As an embodiment using a conventional technology, a domestic patent (application number 10-2000-0070425, publication number 2002-040387, name; IP message exchange device and method using an Ethernet switching device) is an Ethernet in an ATM (Asynchronous Transfer Mode) switch. An apparatus and method for control period communication (IPC) using a switching device. It is an object of the present invention to provide an IPC message exchange method using an Ethernet switching scheme in which an Ethernet switching device is embedded in an ATM switch main controller module and a radial Ethernet network dedicated to IPC messages is constructed using a backplane for matching each module.

In the case of the prior invention, Ethernet switch 105 for IPC has a fatal effect on routing function transfer in case of failure. IPC path redundancy may be considered. In addition, it is impossible to determine the capacity indefinitely due to the influence of the capacity of the Ethernet switch 105 and the number of ports provided for the IPC. IPC path redundancy can lead to an increase in failure factors as well as a price increase in equipment development, and has a fundamental problem in scalability.

A scalable architecture allows you to easily change capacity based on your service offering capacity on a single platform, enabling scalability from edge level to backbone core level and all nodes managed in one central table. It means a scalable router. That is, the prior art had a problem of limitation of expansion in this respect.

Therefore, the technical problem to be achieved by the present invention is to solve the problems of the prior art as described above, by configuring the IPC configuration in the form of a packet (Packet Ring) packeting configuration to enable IPC transmission without an Ethernet switch The purpose of the present invention is to provide an IPC message exchange device and method.

That is, the technical problem to be achieved by the present invention is to configure the IPC configuration in the form of a packet ring, so that the IPC message can be easily transmitted using the ring restoration function and the path change, and the expansion is easy. It is an object of the present invention to provide an IPC message exchange apparatus and method using a packeting configuration that can provide a scalable router configuration.

Another object of the present invention is to provide an IPC message exchange apparatus and method using a packeting configuration that can increase the stability of the system and further simplify the system configuration, thereby reducing the overall cost.

1 is a configuration of a network device according to the prior art and an embodiment of the IPC using the Ethernet switch.

2 is a block diagram of an IPC message exchange apparatus using the packeting configuration of the present invention.

3 is a block diagram of an IPC packeting MAC (MAC) device for explaining the operation of the apparatus and method of the present invention.

<Description of the symbols for the main parts of the drawings>

201 ... main processor

203, 211 ... IPC packeting MAC (Media Access Control) device

205 ... Packing

207 Line Card

208 ... network processor

210 ... Control Processor (CP)

213 ... switch fabric

301, 302, 303 ... buffers

304 ... Packeting Mac Controller

305.Tx Arbiter

306 ... Rx Arbiter

308 Processor Interface Controller

309 ... DMA controller

In order to achieve the above object, the IPC message exchange apparatus using the packeting configuration according to the present invention is configured by configuring the IPC configuration in the form of a packet ring in a distributed system such as a router system or a switch system. It features.

In a preferred embodiment of the present invention, the exchange device comprises: two main processors for redundancy; Two packetings for control period communication (IPC) delivery; It includes two line cards that are connected to each of the two packetings to exchange IPC messages, and through the packeting, message exchange is performed by inserting predetermined information such as routing information into the IPC message.

In a preferred embodiment of the present invention, data forwarding at the exchange is via an interface between a network processor and a switch fabric in the system.

In a preferred embodiment of the present invention, each main processor includes an IPC packeting Media Access Control (MAC) device for the packeting and matching; Each line card includes a host processor for performing the system initialization, adding entries to a predetermined forwarding table, managing multicast groups, and performing management and control functions for changing thresholds in buffer management; The host processor has an IPC packeting MAC device for matching with the packeting; The IPC packeting MAC device receives a buffer as three buffers for controlling the corresponding data flow by determining whether to use this message for the incoming IPC message or to transmit it for use by an adjacent main processor or host processor. And a transmit buffer and a transient buffer.

In a preferred embodiment of the present invention, a transmit arbiter (Tx Arbiter) and a receive arbiter (Rx Arbiter) respectively control the direction and flow of a packet to control the three buffers; A predetermined packeting MAC controller controls a buffer for each IPC packeting MAC device and each arbiter; The packeting MAC controller is matched to the main processor or host processor via a predetermined processor interface controller; The data stored in the receive buffer and the transmit buffer are transferred to a memory of the main processor and the host processor through a predetermined direct memory access (DMA) controller.

And, in order to achieve the above object, the method according to the present invention, in a method for Inter Processor Communication (IPC) in a distributed system, if a predetermined IPC message flows into the IPC packeting MAC device Determining, by the receiving arbiter, whether the message is sent to the user or not based on the unique address of the IPC message; Entering the receive buffer if the IPC message is on the MAC device; Moving to a transient buffer if the destination of the IPC message is not the MAC device or is a message that must be path-through; The received IPC message is moved to a memory in a main processor or a host processor through a DMA controller under the control of a packeting MAC controller, and the moved message carries out respective message processing processing of the main processor or the host processor. Performing; When the IPC message is transmitted, inputting the IPC message generated by the main processor or the host processor into a transmission buffer through a DMA controller in each memory area; And the IPC message to be transmitted controls the packeting MAC controller through the control of the main processor or the host processor through a processor interface controller, which controls the transmit buffer again, in which case the transmit arbiter sees the unique address of the transmitted message. And then sending the IPC message to IPC packeting.

Hereinafter, preferred embodiments of an IPC message exchange apparatus and method using a packeting configuration according to the present invention will be described in detail with reference to the accompanying drawings. In the following description of the present invention, when it is determined that detailed descriptions of related well-known technologies or configurations may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or an operator. Therefore, the definition should be made based on the contents throughout the specification.

In addition, in describing the present invention, the same member number is used for components having the same or similar function, and a lower case alphabet is added after the member number to distinguish corresponding components. Therefore, in the case where only the member number is used without the lowercase alphabet, all the corresponding constituent members shown in the drawings will be indicated.

2 is a configuration diagram of an IPC message exchange apparatus using a packeting configuration according to the present invention, and FIG. 3 is a configuration diagram of an IPC packeting MAC apparatus for explaining an IPC message exchange method using a packeting configuration according to the present invention. .

Referring to FIG. 2, an IPC message exchange apparatus using a packeting configuration according to the present invention configures an IPC configuration in the form of a packet ring to enable IPC message exchange without an Ethernet switch. The apparatus 200 according to the present invention is connected to two main processors 201a and 201b for redundancy, two packetrings 205a and 205b for IPC delivery, and the packetings 205a and 205b for an IPC message. It comprises a line card 207 which can be exchanged.

The apparatus 200 of the present invention exchanges messages by inserting routing information into an IPC message through the packetings 205a and 205b described above, and data forwarding is performed between the network processors (NPs) 208 and the switch fabric 213. Achieved via interface 212.

The MPs 201a and 201b include IPC packeting MAC (Media Access Control) devices 203a and 203b for matching with the packetlings 205a and 205b. Each line card 207 has a host processor (CP) that performs system initialization, adds entries to the forwarding table, manages multicast groups, and manages and controls functions such as changing thresholds in buffer management ( 210, and the CP 210 includes an IPC packeting MAC device 211 for matching with packeting 205a and 205b.

Referring to FIG. 3, the IPC packeting MAC apparatus 203, 211 controls whether the message is to be used by itself or forwarded by an adjacent MP or CP for the incoming IPC message and controls the corresponding data flow. . To this end, the IPC packeting MAC devices 203 and 211 are composed of three buffers: a reception buffer 301, a transmission buffer 302, and a transient buffer 303. In order to control the three buffers, the Tx Arbiter 305 and the Rx Arbiter 306 respectively control the direction and flow of the packet. A packeting MAC controller 304 controls the buffers 302, 301 and 303 and the arbiters 305 and 306 of the IPC packeting MAC devices 203 and 211. Packeting MAC controller 304 is matched to MPs 201a and 201b or CPs 210 via processor interface controller 308, respectively. The data stored in the buffers 301 and 302 are transferred to the memory of the MP and the CP through the DMA controller 309 when the data is transmitted and received.

The present invention configured as described above is similar to the register-insertion ring method used in the ring network of the fast subscriber access network. However, the high speed subscriber access network aims to provide the most economical means of delivery from low speed to high speed regardless of the services of the upper layer. Recently, Internet service is a connectionless service, which has a characteristic of dynamically changing the path depending on the traffic conditions of the network. However, if a link or node failure occurs, it can have serious consequences for the network. The development of the Resilient Packet Ring has been recently made. Since the present invention is not related to packetization of a fast subscriber access network, but is related to a packeting configuration inside a router system, there is a difference from the RPR technology used in a conventional fast subscriber access network.

Referring to FIG. 3, when the IPC message 325 flows into the IPC packeting MAC device 203 (211), the receiving arbiter 306 determines whether the message is sent to itself or not based on the unique address of the IPC message. do. If the message came to itself, it enters the receive buffer 301. On the other hand, if the destination of the message is a different place or a message that needs to be passed through, it is moved to the transient buffer 303. In this case, it moves in the direction of the IPC packeting MAC device 324, 320, 321, 303 in the next MP or CP module.

The received IPC message is moved 317 to the memory in the MP or CP via the DMA controller 309 under the control of the packeting MAC controller 304. The moved message performs the respective message processing processing of the MP or CP.

When transmitting an IPC message, the IPC message generated by the MP or the CP is input to the transmission buffer 302 via the DMA controller 309 in each memory area (316). The IPC message to be transmitted controls the packeting MAC controller 304 via the processor interface controller 308 to control the MP or CP, which in turn controls 310 the transmit buffer. In this case, the transmission arbiter 305 selects the destination direction based on the unique address of the transmission message and then sends the IPC message to the IPC packeting 205a and 205b.

As described above, the present invention provides an IPC configuration in the form of a packet ring to enable IPC transmission without an Ethernet switch.

That is, according to the present invention, if the IPC configuration is configured in the form of a packet ring, the IPC message can be easily transmitted by using the ring restoration function and the path change, and the expansion is easy. Configuration can be provided.

As a result, the present invention has the effect of increasing the stability of the system, and further simplify the system configuration based on the above-described advantages can reduce the overall cost.

Although a preferred embodiment of the present invention has been described in detail above, those skilled in the art to which the present invention pertains may make various changes without departing from the spirit and scope of the invention as defined in the appended claims. It will be appreciated that modifications or variations may be made. Therefore, changes in the future embodiments of the present invention will not be able to escape the technology of the present invention.

Claims (5)

An apparatus for Inter Processor Communication (IPC) in a distributed system, the apparatus comprising: Two main processors for redundancy; Two packetings for control period communication (IPC) delivery; Two line cards each connected to the two packetings to exchange IPC messages, IPC message exchange apparatus using a packeting configuration, characterized in that the exchange of messages by putting predetermined information such as routing information into the IPC message through the packeting. 2. The apparatus of claim 1, wherein data forwarding in the switching device is performed through an interface between a network processor and a switch fabric in the system. 2. The apparatus of claim 1, wherein each main processor includes an IPC packeting Media Access Control (MAC) device for packeting and matching; Each line card includes a host processor for performing the system initialization, adding entries to a predetermined forwarding table, managing multicast groups, and performing management and control functions for changing thresholds in buffer management; The host processor has an IPC packeting MAC device for matching with the packeting; The IPC packeting MAC device receives a buffer as three buffers to control the corresponding data flow by determining whether to use the message for the incoming IPC message or to be used by an adjacent main processor or host processor. And an IPC message exchange apparatus using a packeting configuration, comprising a transmit buffer and a transient buffer. 4. The apparatus of claim 3, wherein a transmit arbiter and a receive arbiter respectively control the direction and flow of packets to control the three buffers; A predetermined packeting MAC controller controls a buffer for each IPC packeting MAC device and each arbiter; The packeting MAC controller is matched to the main processor or host processor via a predetermined processor interface controller; IPC using a packeting configuration, wherein the data stored in the reception buffer and the transmission buffer are moved when data is transmitted and received in a memory of the main processor and the host processor through a predetermined direct memory access (DMA) controller. Message exchanger. A method for Inter Processor Communication (IPC) in a distributed system, When a predetermined IPC message flows into the IPC packeting MAC device, determining whether the received arbiter is based on a unique address of the IPC message to determine whether the message is sent to itself; Entering the receive buffer if the IPC message is on the MAC device; Moving to a transient buffer if the destination of the IPC message is not the MAC device or is a message that must be path-through; The received IPC message is moved to a memory in a main processor or a host processor through a DMA controller under the control of a packeting MAC controller, and the moved message carries out respective message processing processing of the main processor or the host processor. Performing; When the IPC message is transmitted, inputting the IPC message generated by the main processor or the host processor into a transmission buffer through a DMA controller in each memory area; The IPC message to be transmitted controls the packeting MAC controller through the control of the main processor or the host processor through a processor interface controller, which controls the transmission buffer again. In this case, the transmission arbiter looks at the unique address of the transmission message and sets the destination direction. And transmitting the IPC message to the IPC packeting after the selection.