WO2004077760A1

WO2004077760A1 - Efficient implementation of simple nodes in a switched network

Info

Publication number: WO2004077760A1
Application number: PCT/EP2004/001174
Authority: WO
Inventors: Joachim Feld; Günter Steindl; Karl Weber
Original assignee: Siemens Aktiegesellschaft
Priority date: 2003-02-28
Filing date: 2004-02-09
Publication date: 2004-09-10
Also published as: DE10309168A1

Abstract

The invention relates to a system, a method, a controlling means (1), and a switching coupling element (2) for transmitting data via a network (4), particularly an Ethernet. The switching coupling element (2) comprises two interfaces (5a,b) for transmitting data and another interface (5c) to the application (3) running on the controlling means (1). The addresses of additional applications (3i) with which communication takes place are managed in a destination routing table (6) on the switching coupling element (2). Addresses are determined via APR cache. Data messages have to be forwarded only in one direction of the network (4) that is configured as a line.

Description

description

Efficient implementation of simple nodes in a switched network

The invention relates to a system, a method, a controller and a switching coupling element for data transmission over a network, in particular an Ethernet.

In automation technology, there is an increasing requirement to use Ethernet for data transmission also at the field level. This use of the Ethernet provides easy access to the Internet. In addition, the use enables a problem-free configuration of the network that is used for data transmission. Linked to the implementation of Ethernet for data transmission at the field level, however, are requirements for simple cabling structures. It must also be possible to implement the networks at low cost. At the same time, using Ethernet at the field level requires that the entire system is extremely robust.

Nowadays, the wiring of the field devices or the controls in automation technology is usually carried out using a linear structure. The advantage here is that it is easiest to expand and the costs for the respective connection points of the field devices are relatively low. However, when using an Ethernet as a data transmission system, there is a requirement for inexpensive switches with a large number of ports

(Interfaces) to use. However, such a structure leads to confusing cabling, since one moves away from a linear structure to a tree-like structure with a complex topology. In addition, switches that are used in such a complex topology require a large amount of memory because a large number of addresses have to be managed and a large amount of memory is required in the case of high message volume for the buffers in which the messages can be stored must be used. For example, a switch, which is positioned between a controller and an extensive manufacturing plant and is also connected to an office network, must store and manage up to 1000 and more Ethernet addresses and handle a large amount of real-time data traffic, which is a high level Storage capacity required so as not to lose any messages.

In order to be able to implement an Ethernet with simple cabling anyway, switches with a small number of ports (interfaces) are used today. The use of these switches leads to relatively high costs because, for example, two ports on a four-port switch must already be used for the transmission of the messages and only two devices or field devices can be connected to the switch. Such a structure is inefficient and expensive because a large number of switches are required.

Alternatively, the switch functionality can be integrated into the end device or the controller. However, this requires the provision of additional memory for the switch functionality (buffer, address table). The provision of the additional storage capacity causes considerable expenditure in the form of costs if the memory is made available as an integrated component on the controller, or a relatively high transmission bandwidth between the controller and the memory if the memory is swapped out.

In order to use less storage space, eliminating the address table would be a possible solution, but this would result in all messages being sent in all directions over the network. However, such a procedure significantly increases the communication load in the system. The invention is based on the object of enabling an efficient implementation of simple nodes in a switched network, in particular an Ethernet, for the data transmission.

The object is achieved by a system for data transmission over a network, in particular an Ethernet, with at least one node designed as a controller, at least one switching coupling element integrated in the controller, the switching coupling element having two interfaces as connections to the network and / or one the controller running application locally connected to the switching coupling element, the controller having first means for recognizing the interface via which data is received and / or second means for assigning the interface via which data is sent, and / or wherein the application is using Has connection information to other applications available in the network and / or the switching coupling element has a direction table for storing entries based on the connection information, the entries specifying the interface via which data is used to an application.

The object is further achieved by a method according to claim 7, a control according to claim 16 and a switching coupling element according to claim 21.

The invention is based on the finding that when using a network, in particular an Ethernet, the simplest and most efficient possible implementation of network components is desirable for data communication in the automation environment. This simple, efficient and secure data transfer must be made possible in a cost-effective manner. For this reason, a network designed as a line is generally used. Here, the field devices that are connected to the network hang in linear Structure, for example on the Ethernet. Forwarding of data, in particular data telegrams, to the corresponding destinations, that is to say special applications on the field devices, must be possible on the network as simply as possible. For this reason, it is advantageous if switching elements, so-called switches, which have only a few interfaces, are used for data transmission. A simple network topology is guaranteed when using such switches with few interfaces (ports). At the same time, the number needed

Switches with only a few interfaces or ports with such a topology are relatively high. For example, a four-port switch can only supply data to two field devices, since the other two ports must be available for the forwarding of data within the network. For this reason, it is advantageous if the switches can be integrated directly into the field devices or controllers. In this case, there is no additional hardware expenditure for the installation of switches. For the implementation of such integrated switches, however, it is a requirement that they require little storage space, since integrated storage space is a cost factor.

In order to be able to integrate switches with low storage space requirements in the field devices or controls, simple addressing must be guaranteed. This means that the switches have an address table in which the Ethernet addresses of all relevant communication participants or applications are entered. Such address tables require a lot of storage space. To reduce this, a direction table for address management is used in the proposed system.

All that is entered in the direction table is whether an application to which a data telegram is to be used is located in the part of the network or the Ethernet or whether the application is in the part of the Ethernet to which the other interface of the switch leads. In this case, the field device or application on the controller only has to specify where the data telegram is to be sent to. The addressee is entered in the direction table on the switch integrated in the controller and the data telegram is only sent in one direction of the network designed as a line via the port or the interface read from the direction table. A direct address that requires significantly more storage space does not have to be specified. The recipient of the data telegram now recognizes when the data telegram arrives that he is selected as the addressee for this data telegram. In this case, it does not send the telegram to the remaining field devices or controllers on the line. The telegram is forwarded by the other participants hanging on the line.

With this procedure you do not have to save all addresses, just enter those in the direction table with which you really communicate. The advantage of such a data transmission is that relatively little storage space has to be used for address management. In addition, the data telegrams are only sent in one direction of the linear network. Unnecessary

Data traffic within the Ethernet can be avoided in this way. Cost-effective and efficient data transmission on the Ethernet is guaranteed.

It is also advantageous that the system is easy to scale. If a new controller is added or integrated into the network, it is not necessary to additionally create a switching option that establishes a branch within the network and supplies the new controller or the applications running on it with data. When installing such a controller that over Data telegrams for the applications are already automatically guaranteed.

A further advantageous embodiment of the invention is characterized in that the use of an address

Resolution protocol cache (ARP cache) is provided to provide the connection information. With the help of this standard technology, when installing the entire network or installing a new controller in the already existing Ethernet, it can be determined where the other applications relevant to the corresponding field device are located within the network, to which data telegrams may be sent Need to become. For example, an application simply sends an ARP request that asks where the desired target application is located. The use of this standard technology within the framework of the system enables the necessary information to be built up easily, which can then be entered in the direction table in the form of connection information. Using this method, an application or a controller can easily obtain information about the direction in which the other relevant applications can be found in the network formed as a line.

Another advantageous embodiment of the system is characterized in that the network is designed as a line. This is advantageous because a line represents a simple topology and new controllers or field devices to be added to the network can be easily integrated. The clarity of both the network and the communication structure is given in this case. It is also advantageous that switches with a few ports can be installed when using a line, which means that, for example, the individual switches are not confronted with too much data that is sent by any number of field devices that are connected to the switch become. A further advantageous embodiment of the invention is characterized in that the switching coupling element is provided for recognizing a data telegram as directed to the application. It is advantageous here that the data telegrams do not have to have an exact destination address, but that they are recognized as belonging to a control or application as soon as they pass the corresponding addressee on the Ethernet. The switches can thus automatically "fish" the data telegrams intended for them from the network. The use of such a methodology leads to an advantageous saving of memory space when assigning addresses.

A further advantageous embodiment of the invention is characterized in that the switching coupling element is provided for forwarding a data telegram that is not directed to the application. Here, data telegrams that have been sent by a transmitter via the Ethernet are forwarded to various addressees in the sense of a multicast. The data telegrams are thus passed on or forwarded by the switches within the network if they are not intended for the application that is implemented on the controller in which the respective integrated switch is located, which the data telegram passes. Automatic forwarding of all data telegrams within the Ethernet is thus easily guaranteed. The telegrams are forwarded until they pass the addressee for whom they are intended. This addressee then “fishes” the data telegrams intended for him from the communication stream.

A further advantageous embodiment of the invention is characterized in that the switching coupling element has an intermediate buffer for storing data messages. This training is particularly advantageous when there is a high volume of data, ie when there is a high production of data telegrams. comes. In this case, the data telegrams can be buffered in the buffer and then, if data transmission is possible again, sent on via the Ethernet. The use of these intermediate buffers ensures that data telegrams or messages are not lost in the context of the system according to the invention.

A further advantageous embodiment of the invention is characterized in that, within the scope of the method, the stored data telegrams are sent via the selected interface when there is a pause in transmission at the interface. It is advantageous here that the data telegrams stored in an intermediate buffer are then introduced or used in the communication system when capacities for the data transmission have become free again. Up to this point in time at which such a transmission pause occurs, the data messages are stored in the intermediate buffer. It is also advantageous here that the data telegrams have already been assigned the direction in which they are to be used via the Ethernet. As soon as a transmission pause occurs, the corresponding telegram can be sent via the corresponding port without further request regarding the destination address. A rapid processing of the data transmission is hereby advantageously guaranteed.

A further advantageous embodiment of the invention is characterized in that, within the scope of the method, the stored data telegrams are assigned a negative acknowledgment if they cannot be sent via the selected interface within a defined period of time. Here, for example, a certain time is defined as part of a timeout for which the data telegrams are to be stored in the buffer. If the data telegram is not sent within this time, the data telegram is informed that it is not within the scope of the is then returned to the corresponding application that wanted to send the data. The application is thus informed that the data transfer has not been carried out. In this case, the application can carry out another data transmission again. It is advantageous here that the information that a data transmission has failed is available within the framework of the system. In this way, loss of information can be prevented.

The invention is described and explained in more detail below with reference to the exemplary embodiment shown in the figure.

The figure shows a schematic representation of the system for data transmission over a network using simple nodes. A switching coupling element 2 or a switch is integrated in a controller 1 or a controller. An application 3 is running on the controller. 4 data are exchanged with other applications 3i via the network. The network 4 can be designed as an Ethernet. The switch or the switching coupling element 2 has interfaces 5. These interfaces 5 are so-called ports. Two interfaces 5 _a , 5 _b ensure data transmission over the network 4. A third local interface 5 _σ ensures the connection of the application 3. The switch also has a direction table 6 and an intermediate buffer 7 for storing data telegrams.

In the described system, integrated two-port switches 2 can be used for data transmission over a network 4 or an Ethernet. The two-port switches 2 are integrated in a controller 1 or a controller. The two-port switches 2 do without a special address table in which Ethernet addresses must be stored or managed. Rather, the address table is the direction in which a data telegram is to be sent via the network 4, ie via which port 5 _a , 5 _b . For this purpose, the application 3 has information on the connections which lead to the controller or the node and which originate from it. The connection information is provided for example via the Address Resolution Protocol cache (ARP cache) within the framework of the system. An application 3 thus knows which other applications 3i are connected to it in an automation system via the network 4 and where these applications 3 are located within the network formed as a line. This connection information can be entered in the direction table 6. The port information (which port must be used for a connection or an application relationship) is also managed on this database.

If an incoming data telegram is not directed to this station, it will be forwarded via the other port as soon as possible. With these properties, a special address table for managing twelve-digit, 6-byte long Ethernet addresses is no longer required, although the system behaves exactly like a multicast table. This leads to a reduction in the memory requirement of 10 to 20 kbytes per switch, since the address table which requires memory space can be replaced by a simple direction table 6.

Due to the linear design of the network 4, conflicts can only arise at a port 5 if something is to be sent from the local interface 5 _{C and} something is to be received from at least one of the ports at the same time. In this case, the local message, which reaches the switch from the application 3 via the interface 5 _σ , is buffered in an intermediate buffer 7. The size or the storage space requirement of the intermediate buffer 7 depends on the data volume of the respective station or application 3 and on the port load. Applied by application 3 Depause occurs at the respective port 5 _a , 5 _b . An existing low-priority transmission order is then submitted during a transmission pause. In this case there was no loss of data.

A time limit can also be set in the system, which should not be exceeded when data telegrams are temporarily stored in the intermediate buffer 7. If such a time limit has expired, the corresponding send request for the data telegram is acknowledged locally with a negative acknowledgment. At least the local station or application 3 thus receives an indication that the communication could not be successfully completed due to overload.

If a high-priority order is put in between, this is done as part of a priority control. The high-priority orders should not exceed a certain proportion of the bandwidth. So many buffers should be provided that messages displaced by high-priority telegrams can be buffered within the minimum time limit. If the buffers are filled and high-priority messages are pending, the incoming low-priority messages are pushed out according to the usual blocking mechanism.

In summary, the invention relates to a system, a method, a controller 1 and a switching coupling element 2 for data transmission via a network 4, in particular an Ethernet. Here, the switching coupling element 2 has two interfaces 5 _a , _b for data transmission and a further interface 5 _C for application 3, which runs on the controller 1. The addresses of further applications 3ι, with which communication is carried out, are managed on the switching coupling element 2 in a direction table 6. The address is determined via the APR cache. The forwarding of

Data telegrams only need to be in one direction as a line

Claims

claims

1. System for data transmission via a network (4), in particular an Ethernet, with - at least one node designed as a controller (1), at least one switching coupling element (2) integrated in the controller (1), the switching coupling element (2) has two interfaces (5 _a , _b ) as connections to the network (4) and / or - an application (3) running on the controller (1) and connected locally to the switching coupling element (2), the controller 1 having first means for Detection of the interface (5 _a , _b ) via which data is received and / or second means for assigning the interface (5 _a , _b ) via which data is sent, and / or wherein the application (3) has connection information to other applications (3χ) present in the network (4) and / or wherein the switching coupling element (2) has a direction table (6) for storing entries (6ι .. _n ) on the basis of the connection information, the entries ( 6ι..n) d Specify the interface (5 _a , _b ) via which data is sent to an application 3 _± .

2. System according to claim 1, which also means that the use of an address resolution protocol cache is provided to provide the connection information.

3. System according to one of claims 1 or 2, characterized in that the network (4) is designed as a line.

4. System according to one of claims 1 to 3, so that the switching coupling element (2) is provided for recognizing a data telegram as directed to the application (3).

5. System according to one of claims 1 to 4, so that the switching coupling element (2) is provided for forwarding a data telegram not directed to the application (3).

6. System according to one of claims 1 to 5, so that the switching coupling element (2) has an intermediate buffer (7) for storing data telegrams.

7. Method for data transmission over a network (4), in particular an Ethernet, in which - at least one node is designed as a controller (1), at least one switching coupling element (2) is integrated in the controller (1), the switching coupling element (2) has two interfaces (5 _a , _b ) as connections to the network (4) and / or - an application (3) running on the controller (1) is locally connected to the switching coupling element (2), the interface ( 5 _a , t _> ) _{r received} via the data and / or via the data sent by the controller (1) and / or wherein the application (3) has connection information to other applications (3i) present in the network (4) ) is provided and / or wherein in the switching coupling element (2) on the basis of the connection Are stored (6), said entries (6ι .. _n) the interface (5 _a, _b) specifying be sent via the data to an application (3i).

8. The method according to claim 7, d a d u r c h g e k e n n z e i c h n e t that for the provision of the connection information

Address Resolution Protocol cache is used.

9. The method according to any one of claims 7 or 8, d a d u r c h g e k e n n z e i c h n e t that the network (4) is designed as a line.

10. The method according to any one of claims 7 to 9, so that the switching coupling element (2) recognizes whether a data telegram is directed to the application (3).

11. The method according to any one of claims 7 to 10, that a data telegram not directed to the application (3) is forwarded by the switching coupling element (2).

12. The method according to any one of claims 7 to 11, so that a transmission priority is assigned to the data telegrams.

13. The method according to any one of claims 7 to 12, characterized in that data telegrams are stored in an intermediate buffer (7) of the switching coupling element (2).

14. The method according to any one of claims 7 to 13, characterized in that the stored data telegrams are sent via the selected interface (5 _a , _b ) when a transmission pause occurs at the interface (5 _a , _b ).

15. The method according to any one of claims 7 to 14, characterized in that the stored data telegrams are assigned a negative receipt if they cannot be sent within a specified period of time via the selected interface (5 _a , _b ).

16. Control for data transmission over a network (4), in particular an Ethernet, with at least one integrated switching coupling element (2), the switching coupling element (2) having two interfaces (5 _a , _b ) as connections to the network (4) and an application (3) running on the controller (1) and locally connected to the switching coupling element (2), the controller (1) receiving first means for recognizing the interface (5 _a , _b ) via the data and / or second means for assigning the interface (5 _a , _b ) via which data is sent, and / or wherein the application (3) has connection information to other applications (3 _± ) present in the network (4) and / or wherein the switching coupling element (2) has a direction table (6) for storing entries (6ι .. _n ) based on the connection information, the entries (6ι .. _n ) specifying the interface (5 _a , _b ) via the data to an application (3i) are sent.

17. The controller as claimed in claim 16, and that the use of an address resolution protocol cache is provided to provide the connection information.

18. Control according to one of the claims 16 or 17, so that the switching coupling element (2) is provided for recognizing a data telegram as directed to the application (3).

19. Control according to one of claims 16 to 18, that the switching coupling element (2) is provided for forwarding a data telegram not directed to the application (3).

20. Control according to one of claims 16 to 19, so that the switching coupling element (2) has an intermediate buffer (7) for storing data telegrams.

21. Switching coupling element for data transmission over a network (4), in particular an Ethernet, which has two interfaces (5 _a ,) as connections to the network (4) and / or a direction table (6) for storing entries based on connection information , with the entries specifying the interface (5 _a , _b ) via which data are sent to.

22. Switching coupling element according to claim 21, so that the switching coupling element (2) has an intermediate buffer (7) for storing data telegrams.