Multiplex communications – Data flow congestion prevention or control – Flow control of data transmission through a network
Reexamination Certificate
1998-09-04
2002-11-05
Chin, Wellington (Department: 2664)
Multiplex communications
Data flow congestion prevention or control
Flow control of data transmission through a network
C370S338000
Reexamination Certificate
active
06477147
ABSTRACT:
BACKGROUND OF THE INVENTION
Local computer networks which are configured according to the IEEE 802.3 Standard, which is referred to below as Ethernet Standard, constitute a technology with which terminals are connected via a commonly used, serial bus. The access to this bus is regulated by means of the so-called Carrier Sense Multiple Access/Collision Detection (CSMA/CD). The Ethernet protocol is a so-called fair protocol, i.e. when there is competing access, all the terminals connected to a bus have, on a statistical average, the same chance of transmitting via the bus their data which is to be transmitted, regardless of the type of data to be transmitted.
However, in many cases it is necessary to prefer certain data flows over others. Examples of this are data streams with real-time requirements, for example a sound data stream or a video data stream, but also alarm messages for controlling machines. Such data streams require a certain minimum quality during their transmission. Guarantees for such minimum quality levels are, however, currently not ensured with the Ethernet protocol, since with the CSMA/CD protocol it is not possible to make a distinction between data streams and therefore different treatment of data streams is also impossible.
The Ethernet bus is a passive transmission medium and the switching work is distributed between the terminals which also operate according to the Ethernet protocol. The so-called Switched Ethernet is a technology in which local networks according to the IEEE 802.3 Standard are coupled by means of packet switching instead of the otherwise customary Ethernet bus. Packet switching is implemented by means of the so-called Ethernet node. An Ethernet node has a plurality of bidirectional accesses, the so-called ports. The Ethernet node must pass on incoming data packets through at least one output port which is coupled directly or indirectly to the destination or destinations of the data packet. If different input packets are received and have to be output via the same output port, the packets are buffered. The buffer used for this can overflow in cases of high load and newly arriving data packets are lost in this case.
Various packet formats of the Ethernet protocol are known from document by U. O. Pabrai, UNIX Internetworking, Artech House, Boston, London, page 23, 1993.
SUMMARY OF THE INVENTION
The invention is based on the problem of specifying a method which is compatible with the Ethernet Standard and with which a certain minimum quality of the transmission for data streams is ensured with respect to real-time requirements for the transmission.
In the method, a priority class is assigned to each data packet to be transmitted by a first arrangement, and the data packet is marked according to the priority class. Then, the data packet is transmitted to a second arrangement taking into account the priority class.
A considerable advantage of the method is especially the fact that it is possible to prioritize the data packets, for example depending on the type of the data stream to be transmitted. In this way, it is possible, on a statistical average, to ensure the required quality for data packets which require a higher priority because of quality requirements relating to real-time requirements during transmission.
The expression “on a statistical average” is to be understood in this context to mean that it is possible to ensure the quality requirements with a certain degree of probability according to the priority class used. The cause of this that the data packets are transmitted according to their priority, for example data packets with a relatively high priority are preferred over data packets with a relatively low priority.
In order to be able to implement a multi-stage communications connection, i.e. a communications connection via several Ethernet switching nodes, it is advantageous in one development of the invention that the data packet is received from a second arrangement, the priority class assigned to the data packet is determined and in the case of which, in turn, the data packet is transmitted on taking into account the priority class.
By means of this procedure, the method is simplified since an entirely new priority class for the data packet does not, in turn, have to be formed in each switching node and assigned to the data packet, rather merely the priority class which has previously been assigned to the data packet is determined, and the priority class continues to be used in the rest of the method.
In order to assign the priority class to the data packet it is advantageous to analyze information contained in the data packet fed to the so-called Ethernet layer and relating to the type of the data packet and thus to the type of the communications connection or else to the type of the data stream, and to take into account the analyzed information in the assignment of the priority class. In this way, it becomes possible automatically to analyze and to ensure the type of the data packet, and thus the quality requirements of the data stream.
In addition, in one development of the invention it is advantageous to carry out an access check for the data packet, as a result of which it is possible to prevent the arrangement with which the method is carried out from being overloaded as a result of an excessive number of data packets to be transmitted.
In addition, in one development of the invention it is advantageous to divide up a buffer of the arrangement into a plurality of parts and to assign each part of the buffer to at least one priority class in each case. This means that in each case only data packets of the corresponding priority class can be stored in the corresponding part of the buffer.
The development described above is improved even more by virtue of the fact that the data packets are output from the corresponding parts of the buffer and transmitted in a predefinable sequence. The sequence can be predefined according to any desired scheduling method.
In addition it is advantageous that, in the event of the buffer memory, or part of the buffer memory in an arrangement which is configured as an Ethernet switching element, overflowing, the data packet causing the overflow is rejected only if it is not possible to abort the reception procedure by means of an artificial collision for this data packet.
In addition, in one development it is advantageous to generate the artificial collisions of the data packets only for data packets for whose source arrangement the generation of an artificial collision is known to be permissible, i.e. can be processed by the transmitting arrangement, for example. This development makes a selection for the arrangements which can process an artificial collision of a transmitted data packet.
In addition, in one development it is advantageous to reserve, during the entire communications connection, resources which are negotiated in a connection setup phase and are required during the communications connection, for data packets which are assigned to a prescribable communications connection. In this way, considerably more reliable guarantees for the quality levels which are required of the data packets to be transmitted are fulfilled.
The arrangement for carrying out the method has a data-packet classification unit, a priority-marking unit and a buffer memory. Using the data-packet classification unit, the data packet is assigned the priority class, and using the priority-marking unit the data packet is unambiguously labeled in accordance with the priority class. The data packets to be transmitted are buffered in the buffer.
In addition, in one development it is advantageous to provide a resource-administration unit in the arrangements and the switching node, which unit has at least one of the following functions:
administration of resources,
access control for the data packet and/or
signaling between the coupled arrangements.
By virtue of the resource-administration unit, a simplified organization of the execution of the method is possible, enabling the method to be improved.
In addition, it is
Böcking Stefan
Braun Bodo
Huth Hans-Peter
Chin Wellington
Pham Brenda
Siemens Aktiengesellschaft
Staas & Halsey , LLP
LandOfFree
Method and device for transmitting a data packet using... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for transmitting a data packet using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for transmitting a data packet using... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2937252