Method for operating a multiport MAC bridge having ports which can be switched off according to an isochronous data stream at one port or port pair in ethernet LANs

This application claims the benefit, under 35 U.S.C. §365 of International Application PCT/EP2009/063333, filed Oct. 13, 2009, which was published in accordance with PCT Article 21(2) on Apr. 22, 2010 in German and which claims the benefit of German patent application No. 10 2008 051 861.1, filed Oct. 16, 2008.

The invention relates to a method for operating a multi-port bridge according to the precharacterizing clause of Claim 1 .

US 2005/0265330 A1 discloses a network relay system which performs a data transmission to and from an opposite device.

The network relay system comprises a number of ports which are connected to corresponding ports of an opposite facility. The network relay system also comprises a connection aggregation control module which aggregates the connections in order to set up a connection aggregation which is considered to be a logical connection and performs a transmission of a control frame signal comprising a synchronization bit and a data transmission to each port via the connection aggregation. As a response to the detection of the occurrence of interference in one of the aggregated connections, the connection aggregation control module causes a port connected to a normal interference-free connection to stop the data transmission and to send the control frame signal which comprises the synchronization bit set to a first value. The first value is used for stopping the data transmission at a corresponding port of the opposite device which is connected to the normal connection.

From U.S. Pat. No. 7,065,050 B1, a method and a device for controlling the data flow in a network switching facility is known. The method comprises the steps of determining whether a set of data arranged in a queue for a port exceed a first predetermined threshold value and the deactivation of the data flow to the port if it is determined that the set of data arranged in the queue have exceeded the first threshold value and the reactivation of the data flow to the port if a predetermined spatial condition and a predetermined temporal condition have been met. The device comprises at least one queue in connection with at least one data port interface for the reception of data which are transmitted to at least one data port interface and a memory management unit in connection with at least one queue. The memory management unit is designed in such a manner that it deactivates the data flow when a filling level of the data in the queue reaches a predetermined threshold value and thereafter reactivates the data flow to the queue when the filling level of the data in the queue reaches a second predetermined threshold value and a predetermined time has elapsed.

WO 2004/068798 A2 describes a method of prioritizing an isochronous data stream which is transmitted from a first bus of a first type of construction having a first bandwidth to a second bus by using a medium having a second bandwidth which is narrower than the first bandwidth.

It is also known to use multi-port bridges in order to divide networks into different layer-1 segments. The associated load in the entire network is considerably reduced since each segment only transports the frames whose receivers are also located in this segment. For real-time applications which demand a guaranteed data rate, prioritization capabilities (traffic classes or traffic types) have been created and listed in the IEEE802.1D standard (Institute of Electrical and Electronic Engineers). In practice, the traditional division of the data streams into different priority classes is not sufficient for guaranteeing the higher demands made on the data throughput.

The object of the present invention then consists in improving the real-time-capability of a data stream in comparison with the prior art and minimizing the propagation delay and the output jitter with respect to the input.

This object is achieved by a method having the features of Claim 1 . Advantageous embodiments of the invention are defined in further claims.

In the method for operating a multi-port bridge for routing differentiated in accordance with traffic classes (meaning, in other words, that the traffic class or information about the traffic class, respectively, are taken into consideration during the routing) of data packets, an additional traffic class is introduced. In one embodiment of the invention, the data packets are subdivided into different priority classes in accordance with the IEEE802.1D standard and an additional traffic class is defined. The multi-port bridge is configured in such a manner that, for each transmission direction, at least one predetermined port is operated in the “forwarding” state and, if necessary, is set into the “forwarding” state in which the predetermined port exclusively forwards data packets of the additional traffic class from a queue. If these are bidirectional data connections, a port pair is set into or operated in the “forwarding” state. In this context, all ports apart from the predetermined port or the predetermined port pair are deactivated during the forwarding.

In one embodiment of the method according to the invention, the method comprises a first checking step in which it is checked whether the additional traffic class is assigned to the arriving traffic. The method is continued in the case of a positive result of the first checking step in a second checking step in which it is checked whether an isochronous data flow is currently active. The method is continued in accordance with the IEEE802.1D standard in the case of a negative result of the first checking step.

According to one embodiment of the method according to the invention, the method, is continued in the case of a positive result of the second checking step in a triggering step in which a watchdog circuit is triggered which, in operation, is used for informing a control unit about a currently active isochronous data stream. The control unit regulates the transmission of the traffic stream in the additional traffic class. In the case of a negative result of the second checking step, in contrast, a subprocess comes into operation which comprises at least one of the following steps:

a first deactivation step in which all ports are switched off, with the exception of the predetermined port or port pair. Thus, only the input and output port of the isochronous data channel is active and, as a result, guarantees the fastest possible transportation without further queue management; a saving step in which the data packets of the queues allocated to the deactivated ports are saved until the respective ports are switched on again; a second deactivation step in which the spanning tree mechanism for the predetermined port or the predetermined port pair is deactivated. In the saving step, all queues of the ports already deactivated are stopped which prevents loss of data when the queues are reactivated again later (the content of the queues is not deleted); an installation step in which a watchdog circuit is installed, the value of which for the timeout time is selected to be of such magnitude that this watchdog circuit can be woken up at any time during the forwarding of the isochronous data stream in the additional traffic class.

The invention has numerous advantages, some of which will be listed in the text which follows:

The queue delay resulting from managing different queues of a number of ports is avoided.

Disturbances of the isochronous data stream are avoided because there are no prioritization mechanisms of other competing ports. This prevents the risk of interrupting the real-time data flow.

The spanning tree mechanisms responsible for avoiding redundant traffic paths via a number of multi-port bridges are bypassed, as it were. This eliminates control frames and the periodic readjustment in the network.

The present invention is not restricted to a predetermined port or predetermined port pair for an isochronous data stream of the new traffic class IC. It is possible to define a number of ports or port pairs for isochronous data transport depending on the data rate specifications of the transmission channel.

FIG. 1 shows a diagrammatic system chart for illustrating an isochronous data flow over two multi-port bridges;

FIG. 2 shows a flowchart of the sequence of the method according to the invention;

FIG. 3 shows a flowchart of the sequence of an IC watchdog process; and

FIG. 4 shows a table which shows the relationship between traffic types, the number of queues, and the user priorities with respect to one another.

FIG. 1 shows a diagrammatic system chart for illustrating an isochronous data flow over two multi-port bridges B 1 , B 2 . The path P of the isochronous data flow leads through a first multi-port bridge B 1 for the routing, differentiated in accordance with traffic classes, of data packets, and a second multi-port bridge B 2 for the routing, differentiated in accordance with traffic classes, of data packets. A first port pair of each multi-port bridge B 1 , B 2 , to which an additional traffic class IC (isochronous channel), newly introduced according to the invention, is statically assigned, can exclusively access the resources (queue). All other ports of the multi-port bridge are switched off. These disconnected ports exhibit the “disabled” status and are accordingly not managed by the bridge management. The queues of the disconnected ports are “frozen” so that the data are not lost after reactivating the respective ports within the queues. Spanning tree mechanisms do not come into operation at the first port pair since the path is exclusively used only by the isochronous data stream.

FIG. 2 shows a flowchart for the sequence of the method according to the invention. The traditional part of the flowchart located outside a rectangular border U is previously known from the IEEE802.1D standard. The traditional part K of the flowchart comprises a frame reception step S 1 in which frames are received, an active topology execution step S 2 following the frame reception step S 1 , in which information about the source port state ZQ and information about the destination port state ZZ are taken into consideration. The topology execution step S 2 is followed by a frame filtering step S 3 which resorts to a filter database.

The traditional part K of the flowchart also comprises a first insertion step S 4 and a second insertion step S 5 in which in each case frames are inserted into queues.

The traditional part K of the flowchart finally comprises a first branch A 1 with the steps, following one another in time, of transmission selection S 6 , priority allocation S 7 , FCS (frame check sequence) recalculation S 8 and frame transmission S 9 .

The innovative part of the flowchart located within the border U, provided with the reference symbol E overall, is related to a novel procedure as provided by the invention. The innovative part E of the flowchart firstly comprises a checking step E 1 in which it is checked whether the additional traffic class IC is assigned to the arriving traffic.

In case of a negative result of the first checking step E 1 , the method is continued in the first insertion step S 4 in accordance with the IEEE802.1D standard. In the case of a positive result of the first checking step E 1 , the method is continued in a second checking step E 2 in which it is checked whether an isochronous data stream is currently active.

In the case of a positive result of the second checking step E 2 , the method is continued in a triggering step E 3 in which a watchdog circuit is triggered which, in operation, is used for informing a control unit that a data stream is currently active. This control unit regulates the transmission of the traffic stream in the additional traffic class.

In the case of a negative result of the second checking step E 2 , a second branch of the flowchart comes into action which comprises the steps mentioned hereinafter, following one another in time:

a first deactivation step E 4 in which all ports are switched off with the exception of the predetermined port or port pair; a saving step E 5 in which the data packets of the queues allocated to the deactivated ports are saved until the respective ports are switched on again; a second deactivation step E 6 in which the spanning tree mechanism for the selected port pair is deactivated; an installation step E 7 in which a watchdog circuit is installed, the value of which for the timeout time is selected to be of such magnitude that this watchdog circuit can be woken up (watchdog triggering) at any time during the flowing of the isochronous data stream.

Both the installation step E 7 and the triggering step E 3 are followed by the first insertion step S 4 located within the traditional part K of the flowchart. An isochronous data stream only requires one queue which is ideally managed in accordance with the FIFO (first in first out) principle. This eliminates the queue management of other queues during the transmission of the isochronous data stream.

FIG. 3 shows a flowchart for the sequence of an IC watchdog process which comprises an inquiry step W 1 in which the watchdog circuit checks whether timeout information is present. The counting process belonging to the watchdog process takes place in block W.

A positive result of the inquiry step W 1 has the consequence that the transmission of the isochronous data stream has been ended (there is no triggering of the watchdog E 3 ). For this reason, all deactivated ports are first switched on again (step W 2 ).

Furthermore, all stopped queues are activated (step W 3 ). In addition, the spanning tree mechanism (step W 4 ) for the ports to which the additional traffic class is not assigned is switched on. After that, the IC watchdog is removed (step W 5 ) after which the multi-port bridge is again in its original state.

FIG. 4 shows a table which documents the relationship between traffic types, the number of queues, and the user priorities with respect to one another.

The table has been modified on the basis of the traditional table known from the IEEE802.1D standard, accommodating an additional traffic class IC.

The abbreviations used in the table have the following significance, in detail:

BK=background, BE=best effort, EE=excellent effort, CL=controlled load, VI=video, VO=voice, NC=network control, IC=isochronous channel.

The table shows that the new traffic class IC prescribes only one queue. No other queues are required so that the queue management is as simple as efficient as possible.