JP2000183935A - Node and intra-node data processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain the transmission of a frame with priority even by an application operating on the basis of a frame without priority or a node using an API. SOLUTION: A node 3 is connected with a network for transmitting a frame with priority through a transmission path 2, and provided with a priority table 12 for relating a port number for specifying a frame to be transmitted to the priority of the frame, and priority information addition processing part 11 for generating a frame with priority by adding priority to a frame without priority received from an application layer 5 by referring to the priority table in an Ethernet driver 8. Thus, the frame becomes the one with priority at the time of being outputted from the transmission path, and the frame without priority can be transmitted and received in upper rank layers than the Ethernet driver.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a node and a data processing method in the node.
The present invention relates to a node used in a data transmission system that transmits and receives frames with priority information and a data processing method in the node.

[0002]

2. Description of the Related Art At present, the most popular LAN (Local Ar
ea Network) There is Ethernet as a system. As is well known, this Ethernet is a method of physically connecting each node (terminal) connected to a network by a star-type wiring and transmitting a packet.
SMA / CD (Carrier Sense Multiple Access with C
The use of the ollision detection method is widespread.

In this CSMA / CD system, a node transmitting a packet checks the use state of the network in advance, sends out the packet if transmission is possible, and confirms that another node is transferring data. After waiting for a certain period, the usage status will be checked again.
Therefore, if the number of nodes connected to the network increases or the amount of data transferred (transmitted / received) between the nodes increases,
The transmission efficiency is reduced (transmission delay time is lengthened) due to waiting for transmission or collision of data (frames) due to lack of bandwidth.
Usually, a network configuration in which each node is connected via a switching hub (concentrator) is mainly used.

This concentrator analyzes the Ethernet head of a frame received at a certain port, and if it recognizes the destination of the frame, relays the received frame only to the port to which the destination node is connected. Like that. Thus, for example, while a frame is being transmitted from the first node to the second node, it is possible to transmit the frame from another third node to the fourth node.
That is, a plurality of nodes can transmit simultaneously.

[0005] However, even in a network using such a concentrator, as a result of a use state check, transmission may wait. The size of the transmission delay time associated with the transmission wait greatly varies depending on the usage status of the network, and furthermore, is determined by the usage status. Is unknown, and the upper limit of the transmission delay time cannot be guaranteed.

Therefore, for example, a LAN in factory automation (FA) requires real-time performance, and a PLC (programmable logic controller)
For frames containing data directly related to control of the production line, such as control systems, the upper limit of the transmission delay time must be guaranteed. It could not be used to construct a LAN for FA.

By the way, data to be transmitted on the FA LAN is directly related to control of the production line, such as communication of a control system terminal (node) such as a PLC, and needs to be transmitted quickly in real time. There is information data that is not directly related to control of the production line and does not need to be transmitted in real time, such as various kinds of statistical information, such as communication between data and a personal computer terminal (node). That is, as described above, a frame transmitted on the network includes a high-priority frame including control system data that needs to be transmitted in real time and a low-priority frame including information-based data that does not need to be transmitted in real time. There is.

[0008] However, in the conventional Ethernet, since frames have no priority information and all frames are treated equally, a frame including control system data which must be transmitted in real time is a frame communication of information system data. May be waited for. The phenomenon is F
The same applies to various LANs, without being limited to the LAN for A.

Therefore, in recent years, a method of providing a field indicating the priority in a frame and transmitting a high-priority frame in preference to a low-priority frame has been studied by IEEE and the like. As a result, the possibility that the communication of the high-priority frame is hindered by the low-priority frame is reduced. Communication becomes possible. That is, it is possible to construct a CSMA / CD-type network using the Ethernet for the FA and the Ethernet.

[0010]

However, the frame transmission system provided with the above-mentioned priority information has the following problems. That is, as described in the embodiment, the concentrator includes a buffer that stores a high-priority frame and a buffer that stores a low-priority frame, and preferentially stores the frame stored in the high-priority buffer. Devices that respond by transmitting data have been developed.

On the other hand, in order to construct a transmission system for frames provided with priority information, it is necessary for the nodes on which the frames are transmitted and received to correspond to the transmission. If so,
Since priority information is added to frames to be transmitted and received, the network I / O A in the BSD version of UNIX
A new extended I / F cannot be used because the SOCKET I / F (socket interface), which is a PI, cannot be used.
It is necessary to formulate (new API for priority).
When a new extended I / F is formulated in such a manner, the existing application must be remodeled to an application capable of transmitting and receiving frames (packets) with priority information corresponding to the extended I / F. The work becomes very complicated.

[0012] Therefore, although it is possible to cope with the concentrator, it is not possible to recreate all the expansion I / Fs and applications for each existing node, and the conventional node is used. Then, since there is no priority information in the frame transmitted from the relevant node, the processing in the concentrator eventually has a low priority, and the above-described conventional frame in which the originally high priority frame cannot be transmitted with priority. Problem is not solved.

[0013] The present invention has been made in view of the above background, and has as its object to solve the above-mentioned problems, without adding a new communication API, and
A node that can transmit and / or receive (communicate) a frame with a priority even if the node uses an application or an API that supports a frame without a priority without rebuilding the application. It is to provide a data processing method.

[0014]

In order to achieve the above-mentioned object, a node according to the present invention specifies a frame to be transmitted, which is a node detachable from a network through which a frame with priority is transmitted. Table (corresponding to “port number” in the embodiment) and the priority of the frame (corresponding to “priority table for each port” in the embodiment). A priority information adding means for generating a frame with a priority by adding a priority with reference to the priority table (claim 1).

The data processing method in a node according to the present invention is a data processing method in a node to be connected to a network, wherein the upper layer has priority information on frames to be transmitted (in the embodiment, (Corresponding to the “priority notification packet”) to the lower layer (in the embodiment, “the layer in which the Ethernet driver is installed”).
The lower layer is a priority table that associates information specifying a frame to be transmitted according to the received priority information with the priority of the frame (corresponding to the “port priority table” in the embodiment). Create Next, a priority-added frame is generated by adding a priority to the non-priority frame received from the upper layer by referring to the priority table (claim 2).

According to the invention of the node and the data processing method described above, a frame without priority received from an upper layer (for example, an application layer) can be converted into a frame with priority by the priority information adding means.
Layers higher than the priority information adding means can transmit in the state of a conventional non-priority frame (packet). Therefore, it is possible to configure a node connected to a network that transmits a frame with a priority without creating an extended I / F or an application capable of processing the frame with a priority in place of a conventional API or the like.

Another solution of the node according to the present invention is a node which can be connected to and disconnected from a network through which a frame with priority is transmitted. Priority information removing means for generating a non-priority frame from which
There is provided a means for sending a frame without priority generated by the priority information removing means to an upper layer.

Another solution of the data processing method in the node according to the present invention is a data processing method in a node to which a network to which a frame with priority is transmitted is connected. When a frame with a priority is received, priority information is removed from the received frame to generate a frame without priority. Then, the generated frame without priority is sent to the upper layer (in the embodiment, a reception notification is made once, and a transfer request from the upper layer is received and sent) (claim 4).

According to the above-described invention of the node and the data processing method, when the data received from the outside (connected network) is a frame with a priority, information relating to the priority is removed from the frame and transmitted to the upper layer. At the time of transmission to the upper layer, at least such priority information is lost, that is, the upper layer can transmit and receive a frame (packet) in a state of a non-priority frame as in the related art.

Therefore, it is possible to configure a node connected to a network for transmitting a frame with a priority without creating an extended I / F or an application capable of processing a frame with a priority in place of a conventional API or the like. .

[0021]

FIG. 1 is a block diagram of an LA to which the present invention is applied.
3 shows the overall schematic configuration of N. As shown in the figure,
A network is constructed by connecting each node 3 via the transmission path 2 with the concentrator 1 at the center. In this network, for example, the node 3 is a LAN arranged in an FA including a PLC, a sensor, a personal computer, and the like. Each node (PLC or the like) connected to the network performs various processes while performing a cooperative operation, a synchronous operation, and the like. Applicable to things. Of course, it is needless to say that the present invention can be applied to various other networks.

The internal structure of the concentrator 1 is as shown in FIG. That is, the concentrator 1 includes a physical layer 1a for transmitting and receiving data and a communication controller 1b. Each physical layer 1a is connected to a predetermined node 3. Thus, the communication controller 1b analyzes the frame received via the physical layer 1a and transfers the frame to the physical layer 1a to which the destination node is connected. Further, in this embodiment, a high-priority buffer 1c for storing high-priority frames and a low-priority buffer 1d for storing low-priority frames are provided. The high priority buffer 1c and the low priority buffer 1
It further includes a flow control function unit 1e that monitors buffer overflow of d.

In the line concentrator 1 having such a configuration, the communication controller 1b detects the priority of the frame sent from the node 3 and registers it in the corresponding buffers 1c and 1d. It has a function of analyzing the contents of the frame stored in the high-priority buffer 1c with priority, and transferring the frame to the physical layer 1a to which the destination node 3 is connected. In addition, the flow control function unit 1e has a function of issuing a transmission suspension instruction to each node when a buffer overflow occurs, for example.

FIG. 3A shows the format of a frame (packet) with priority used in the present embodiment. That is, there is an Ethernet header at the beginning of the frame, followed by Ethernet data which is a data portion to be actually transmitted. The Ethernet header stores DA (destination address), SA (source address), TH (priority information), and L / T (data length) in this order. Further, the priority information includes U
When P is "7", it means high priority, and UP is "1".
Means low priority. Therefore, the type of the priority of the frame can be determined from the UP field. TPID, CFI, and VID are a type ID (fixed to 8100) for specifying a protocol type of an upper layer, routing information (fixed to 0), a virtual LAN, respectively.
ID number (fixed to 0).

The data format of the conventional frame without priority is as shown in FIG.
As is clear from comparison with the frame with priority (FIG. 10A), there is no TH (priority information). In other words, a frame with a priority is obtained by inserting the TH area between the SA area and the L / T area of the frame without the priority.

The node 3 of the present embodiment transmits and receives frames without priorities as shown in FIG. 3B inside (in each layer) of the node 3, and transmits and receives between the nodes 3, that is, the node 3
And the concentrator 1 transmit frames with priority. That is, a function of adding priority information to a frame having no priority information when transmitting to the outside inside the node 3, and removing the priority information from the received frame with priority, and providing no priority. It has a function to convert to frames. Specifically, it is as shown in FIG.

That is, as is well known, the layered structure of Ethernet (PCT / IP protocol) has a layer interface below the application layer 5 as the uppermost layer, as shown in FIG. A transport layer (TCP / IP, UDP) 7, an Internet layer (Ethernet driver) 8, an OS 9, and a physical layer (hardware) 10 are arranged in this order via an API 6 called a socket.

As a data flow in such a configuration, a communication application is created using an API 6 usually called a socket. The data passed from the application to the TCP / IP 7 through the API 6 is further passed to the Ethernet driver 8, and the driver 8 accesses the hardware (Ethernet controller) 10 through the OS 9 to transmit and receive packets to and from the transmission path. Then, the Ethernet driver 8 adds and removes the above-mentioned priority information.

That is, as shown in the figure, in the Ethernet driver 8, a priority information addition processing unit 11 and a priority setting table 12 having a function of executing the flowchart of FIG. 5 and a function of executing the flowchart of FIG. 6 are provided. The provided priority information removal processing unit 13 is provided. A more specific internal structure of the Ethernet driver 8 is as shown in FIG.

As shown in FIG. 7, when an initialization request is received from the upper layer, the driver and the Ethernet board are initialized by the driver initialization processing section 8a and the board initialization processing section 8b. In the transmission request, a packet passed from the upper layer is formed into a packet for Ethernet, and a function of transmitting the packet in order of priority is executed.

The transmission processing unit 8c executes packet transmission to the Ethernet board in accordance with the received transmission request (normal transmission packet (not a priority notification packet from the upper layer)). Then, the priority information addition processing unit 11 of the present invention is arranged in cooperation with the transmission processing unit 8c, and when a priority notification packet is received from an upper layer as described later, the priority information addition processing unit 11 11 operates and stores predetermined data in the priority setting table 12.

In this embodiment, the priority information addition processing section 11 includes a transmission packet analysis processing section 11a for analyzing whether a packet passed from an upper layer is a control packet or a transmission packet, and a priority setting table. A priority table setting processing unit 11b for storing the priority for each port, and searching for the priority of the designated port or manipulating the priority table to convert a packet without priority into a packet format with priority. A transmission packet conversion process 11c is provided.

In the interrupt handler, a receiving packet conversion processing unit 13a as a priority information removing processing unit 13 is provided on the receiving side from the hardware, and the received packet is converted from a packet format with priority into a normal packet format. To the packet format. Further, in the received packet analysis processing unit 13b,
The type of the received packet (whether or not there is a priority) is determined.
a is operated. Further, the received packet analysis processing unit 13b has a function of notifying a higher layer of the reception of a packet having no priority. This notification feature
It also has a conventional one. However, the difference is that in the conventional case, all received packets have no priority, so it should have been notified when received, but in this embodiment, both packets with and without priority may be received Therefore, if there is no priority, the notification is made as it is, and if there is a priority, the notification after conversion is made without priority.

Further, upon receiving a received packet transfer request issued by the upper layer according to the notification, the function of transferring the designated received packet to the upper layer, and upon receiving a termination request from the upper layer, the Ethernet board It has a function to stop and release all resources held by the driver, for example. The functions other than the function of adding / removing the priority information are the same as those of the conventional Ethernet driver, and the detailed description thereof will be omitted.

Here, in the present invention, a special packet format (see FIG. 8) that allows the application 5 to register the priority in the priority setting table 6 provided in the Ethernet driver 8 using the normal API 6 for data transmission is used. Define and Ethernet driver 8 (priority information processing unit 5)
Performs a process of assigning a priority to each packet by referring to the priority setting table.

First, a description will be given of a priority setting packet format passed from the application to the Ethernet driver. In the case of a priority setting packet, as shown in FIG.
Port is 9. Therefore, in cases other than 9, it can be determined that the packet is a normal data packet. Request No. at the head of the data is “1” indicating the setting, and Port No. indicates a port number to which the application attempts to perform the communication with priority (socket API).
In, communication is performed by designating a port number, so that port number is written.)
The information indicating whether the data to be communicated has a high priority (7) or a low priority (1) is set in the "ity" area.

The priority information addition processing section 11 of the Ethernet driver 8 creates the priority table 12 (see FIG. 9) in response to the reception of the above-mentioned priority notification packet, and converts the normal packet (transmission packet). When the packet is received, a packet to which priority information is added is created with reference to the priority table 12, and data is transmitted.

As shown in FIG. 5, first, a transmission request passed from an upper layer is waited for (ST1), and if accepted, the packet type of the transmission request is determined (ST2). That is, it is determined whether or not the packet is transmitted from the ninth port. If it is the ninth port, it can be determined that the packet is for setting the priority, and if not, it is determined that the packet is for data transmission.

If the packet is a priority setting packet, the process jumps to step 9 and sets the priority and the corresponding port in the priority table. Specifically, the data portion of the priority notification packet shown in FIG. 8 is analyzed, and a port number for which a priority is to be set from the Port No area (used when the application transmits a frame of the priority to be set). Port number) and the numerical value stored in the Priority area. Then, the acquired two values (port number and priority) are stored in the port priority table shown in FIG. 9A in association with each other. That is, for example, when a priority notification packet as shown in FIG. 10 is received, the priority of the packet transmitted using the port number 1048 is 7 (high priority) from the data portion, As shown in a), the priority of the corresponding port number is set to 7. After performing such processing, the pre-processing (table creation) for setting the priority ends.

On the other hand, if it is determined in step 2 that the transmission request passed from the upper layer is a packet for data transmission, the process jumps to step 3 to determine whether the packet is a packet divided in the upper layer. Whether the packet is a divided packet can be determined from a flag (not shown) of the IP header.

If not divided, step 6
Then, the priority setting table (priority table for each port) 12 is accessed to read the priority associated with the corresponding port (ST6). That is, for example, if the port number is “1048” registered earlier, the information with the priority “7” is read.

Next, a frame is created in which a TH area (priority information) is inserted after the SA area (source address) of the Ethernet header (no priority information as shown in FIG. 3B) of the received packet. (Convert to packet format with priority). That is, as shown in FIG. 3 (a), only the UP area of the TH area is variable (1/7) and fixed other than that, so the priority (1/7) read from the table and the fixed Each value is inserted in an appropriate order. After that, the data in the format with the priority generated by the conversion is transmitted, and the process ends (ST8).

If the received packet is a fragmented packet, the process jumps from step 3 to step 4, and
Further, it is determined whether or not the packet is the first fragmented packet. This determination can also be made from a flag (not shown) of the IP header. That is, as an actual process, the determinations in steps 3 and 4 can be performed together.

Then, in the case of the first divided packet,
A table in which the ID in the packet is read and the ID is associated with the priority: a priority table for each ID field (FIG. 9)
(See (b)). The priority can be recognized by referring to the port-specific priority table rather than the port number. In other words, since the maximum data length that can be transmitted at one time is determined, when transmitting more data, the data is divided into a plurality of packets and transmitted. At this time, since the second and subsequent packets do not include information for specifying the TCP or UDP source port, the priority cannot be specified only by the information of the second and subsequent intermediate packets.

Therefore, a priority table for each ID field (FIG. 9 (b)) in which the ID number (ID number) of the divided packet has the same value and the ID number of the first packet is associated with the priority. ), The second and subsequent packets can know the priority (1/7) corresponding to the ID number by referring to the priority table for each ID field from the ID number. The acquired corresponding priority can be set as the priority at the time of transmitting the packet.

When the table is created by executing step 5, the process returns to step 6 to execute the above-described processes in order, and to transmit the converted data into a packet with priority. That is, the priority (1/7) corresponding to the port number is acquired by referring to the table (priority data table for each port), and is converted into a packet with priority based on the priority. Since the priority information is already known when executing step 5, step 5 may be skipped to step 7, and a packet with priority may be assembled using the known priority information.

Further, in the case of the second and subsequent divided packets, the result of step 4 is No, so the process directly jumps to step 6 and reads the priority of the corresponding ID from the priority table for each ID field. The packet is converted into a priority-added packet format (ST7) and transmitted to the network (ST8).

The relationship between the above steps and the processing units 11a to 11c shown in FIG. 7 is as follows. The transmission packet analysis processing unit 11a performs the processing of steps 1 to 4 and the priority table setting processing unit 11b executes The transmission packet conversion processing unit 11c performs the processing of steps 5 and 9.
Is performed. The data transmission in step 8 is performed by the transmission processing unit 8c.

As described above, in the present embodiment, two types of tables, a “priority table for each port” and a “priority table for each ID field”, are prepared as the priority table 12, and the data length is obtained by using them. It can handle data to be transmitted after being divided into a plurality of long packets. However, for example, when applied to a network that controls a PLC for FA, since data having a long data length is rarely transmitted so that it must be divided, the priority table for each ID field (FIG. 9B) is It is not necessary. In that case, the processing of steps 3 to 5 shown in FIG. 5 is eliminated.

If a node using an Ethernet driver that does not have a priority table for each ID field transmits a plurality of divided packets,
Since the second and subsequent packets do not have priority information, they can be dealt with by setting a low priority set as an offset, for example. In particular, in the case of a PLC LAN, the fact that the data length is long can be predicted to be an information system in many cases. Therefore, there is no problem in setting such a low priority. Conversely, when a packet having a long data length and transmitted after being divided into a plurality of packets has many high-priority packets, the second and subsequent divided packets may be set to have a high priority. Good.

On the other hand, since the API 6 of the node 3 according to the present embodiment cannot transmit / receive a packet with a priority, the API driver 6 converts a packet with a priority received via a network into a packet without a priority by the Ethernet driver 8. It has become. Furthermore, in this embodiment, in order to be able to cope with a network transmitting and receiving packets without priority,
When a packet without priority is received, it is sent to the upper layer without performing the above-described conversion (normal data processing is of course performed). Specifically, it has a function of executing a flowchart as shown in FIG.

That is, first, if there is an interrupt request from the hardware 10 (interrupt due to reception of a packet sent from the network via the transmission path 2), the received packet is given priority or priority. It is determined whether or not there is no degree (ST11, ST12). This determination is made by the received packet analysis processing unit 13b.

If it is determined that the packet is a priority-added packet, the flow advances to step 13 to convert the packet format with the priority into a normal packet format. That is, the TH area of the Ethernet header of the format with priority (see FIG. 3A) is removed, and the SA area is removed.
The frame is reconstructed into a frame of a 14-byte non-priority format (see FIG. 3B) consisting of an L / T area following the area. This process is executed by the received packet conversion processing unit 13a. Thereafter, the reception is notified to the upper layer, and the process ends (ST14). This processing is performed by the received packet analysis processing unit 13b.

On the other hand, if a packet without priority is received, the process jumps from step 12 to step 14 to notify the upper layer of the reception of the packet and terminate the processing. That is,
Normal reception processing that has been performed conventionally will be executed. This series of processing is performed by the reception packet analysis processing unit 13.
b does.

Normally, there is a request to transfer a received packet from the upper layer according to this notification, so that the received packet is transferred to the upper layer in response to the request. When a termination request is received from an upper layer, termination processing, that is, Ethe
Stop the net board and release all resources held by the driver.

[0056]

As described above, the first invention of the node and the data processing method in the node according to the present invention (claim 1,
In 2), the non-priority frame sent from the upper layer by the priority information adding means is converted (generated) into a frame with priority, so that the upper layer than the priority information adding means has no priority. Transmission and reception can be performed in the state of a frame (packet). Therefore, even if a node communicates using a conventional application, a normal API, or UDP without adding a new API (extended I / F) and without largely changing an application. , A frame with priority can be transmitted to the connected network.

In the second invention (claims 3 and 4),
Since the frame with priority received from the outside by the priority information removing unit is recreated as a frame without priority, a layer higher than the installation layer of the priority information removing unit can receive the frame without priority. Therefore, the first
Similarly to the invention of the first embodiment, there is no need to manufacture an extended I / F or an application that enables communication of a frame with a priority, and even if the node has an API or an application that handles a conventional frame without a priority, A frame with priority can be received and desired processing can be performed.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a network system to which the present invention is applied.

FIG. 2 is a diagram showing an internal structure of the line concentrator.

FIG. 3A illustrates a frame format with priority. (B) is a diagram showing a frame format without priority.

FIG. 4 is a diagram showing a preferred embodiment of a node according to the present invention.

FIG. 5 is a flowchart mainly showing functions of a priority information addition processing unit;

FIG. 6 is a flowchart mainly showing the function of a priority information removal processing unit;

FIG. 7 is a diagram showing the internal structure of the Ethernet driver.

FIG. 8 is a diagram showing a format of a priority setting notification packet sent from an application.

FIG. 9A is a data structure diagram illustrating a port-by-port priority table. (B) is a data structure diagram showing a priority table for each ID field.

FIG. 10 illustrates an example of a priority setting notification packet.

[Explanation of symbols]

 3 Node 8 Ethernet driver (lower layer) 11 Priority information addition processing unit 12 Priority table 13 Priority information deletion processing unit

Claims (4)

[Claims] 1. A priority table in which information identifying a frame to be transmitted and a priority of the frame are associated with each other. And a priority information adding means for generating a frame with priority by adding a priority with reference to the priority table. 2. A data processing method in a node to be connected to a network, wherein an upper layer sends priority information on a frame to be transmitted to a lower layer, wherein the lower layer receives the received priority information. A priority table is created in which information for identifying a frame to be transmitted is associated with the priority of the frame according to the priority table. A data processing method in a node, wherein a frame with a priority is generated by adding a degree. 3. Priority information that is a node detachable from a network through which a priority frame is transmitted, and that generates a non-priority frame by removing priority information from a priority frame received from outside. A node comprising: a removing unit; and a unit that sends a non-priority frame generated by the priority information removing unit to an upper layer. 4. A data processing method in a node to be connected to a network to which a frame with priority is transmitted, wherein when a frame with priority is received from outside, priority information is removed from the received frame. A non-priority frame, and then transmitting the generated non-priority frame to an upper layer.