DETECTION OF A FAULTY NODE IN A NETWORK

• US 20170019290A1
• Filed: 03/19/2015
• Published: 01/19/2017
• Est. Priority Date: 03/31/2014
• Status: Active Grant

• Current Assignee: Siemens AG
• Original Assignee: Siemens AG
• Inventors: Armbruster, Michael, Fiege, Ludger, Riedl, Johannes, Schmid, Thomas, Zirkler, Andreas
• Granted Patent: US 10,103,933 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: H04L 41/0631 using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
• CPC Class Codes: H04L 41/0645 by additionally acting on or stimulating the network after receiving notifications
• CPC Class Codes: H04L 41/0677 Localisation of faults
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 43/0888 Throughput
• CPC Class Codes: H04L 43/091 Measuring contribution of individual network components to actual service level
• CPC Class Codes: H04L 43/16 Threshold monitoring

Abstract

Methods and apparatuses for increasing quality of service in a network having nodes if there is a faulty node which can result in data traffic being overloaded in at least one part of the network are provided. The disclosed embodiments of the invention can be used in the field of safety-critical applications, such as medial applications, monitoring devices, and in-vehicle communication systems.

Claims

1. A method for increasing quality of service in a network having a plurality of nodes if there is a faulty node, in which the nodes are connected to one another via respective connections for the purpose of interchanging data packets, the faulty node is coupled to at least one device, and the at least one device operates as a data sink and/or a data source, having the following steps of:

a) selecting at least one of the nodes as a monitored node;

b) producing at least two observer nodes (KB1, KB2) by selecting from the plurality of nodes, the monitored node being excluded from the selection, in such a manner that both incoming data traffic of a respective data packet class to the monitored node from at least one of the at least two observer nodes and outgoing data traffic of the respective data packet class from the monitored node to at least one of the at least two observer nodes are completely determined;

c) recording the incoming data traffic and the outgoing data traffic of the respective observer nodes;

d) generating expected total outgoing data traffic of the monitored node on the basis of (i) the outgoing data traffic produced by the respective incoming data traffic of the respective data packet class and (ii) expected data traffic of the data source of the at least one device;

e) generating a difference value from a difference between the outgoing data traffic and the expected total outgoing data traffic; and

f) detecting the monitored node as a faulty node if the difference value exceeds a predefinable threshold value.

2. The method as claimed in claim 1, wherein the expected total outgoing data traffic of the monitored node is formed by summing (i) one or more expected outgoing data traffic items and (ii) the expected data traffic of the data source of the at least one device, a respective expected outgoing data traffic item being formed by multiplying (a) a number of outgoing data packets for each incoming data packet of the respective data packet class and (b) the respectively associated incoming data traffic.

3. The method as claimed in claim 1, wherein the data packet class is determined by at least one of the following properties of the respective data packet of the incoming data traffic and of the outgoing data traffic:

e) “

unicast”

forwarding type f) “

multicast”

forwarding type g) “

broadcast”

forwarding type; and

h) priority class.

4. The method as claimed in claim 3, wherein the expected outgoing data traffic is set to be equal to the incoming data traffic for the data packet class of the “

unicast”

forwarding type.

5. The method as claimed in claim 3, wherein the expected outgoing data traffic for the data packet class of the “

multicast”

forwarding type is determined by a result of multiplying an available number of connection outputs of the monitored node to directly adjacent nodes of the monitored node by the incoming data traffic, the available number being determined between zero and a number of connection outputs to directly adjacent nodes of the monitored node which has been reduced by one.

6. The method as claimed in claim 3, wherein the expected outgoing data traffic for the data packet class of the “

broadcast”

forwarding type is determined by a result of multiplying a number of connection outputs to directly adjacent nodes of the monitored node which has been reduced by one by the incoming data traffic.

7. The method as claimed in claim 1, wherein the steps in claim 1 are carried out only if the outgoing data traffic exceeds a predefinable volume of data per unit of time.

8. The method as claimed in claim 1, wherein at least one of the connections of the monitored node, in particular one of the connections going out from the monitored node, is interrupted if the monitored node is detected as a faulty node.

9. An apparatus for increasing quality of service in a network having a plurality of nodes if there is a faulty node, in which the nodes are connected to one another via respective connections for the purpose of interchanging data packets, the faulty node is coupled to at least one device and the at least one device operates as a data sink and/or a data source, having the following units:

a) a first unit for selecting at least one of the nodes as a monitored node;

b) a second unit for producing at least two observer nodes by selecting from the plurality of nodes, the monitored node being excluded from the selection, in such a manner that both incoming data traffic of a respective data packet class to the monitored node from at least one of the at least two observer nodes and outgoing data traffic of the respective data packet class from the monitored node to at least one of the at least two observer nodes are completely determined;

c) a third unit for recording the incoming data traffic and the outgoing data traffic of the respective observer nodes;

d) a fourth unit (M4) for generating expected total outgoing data traffic of the monitored node on the basis of (i) the outgoing data traffic produced by the respective incoming data traffic of the respective data packet class and (ii) expected data traffic of the data source of the at least one device;

e) a fifth unit for generating a difference value from a difference between the outgoing data traffic and the expected total outgoing data traffic; and

f) a sixth unit for detecting the monitored node as a faulty node if the difference value exceeds a predefinable threshold value.

10. The apparatus as claimed in claim 9, wherein a seventh unit which is configured in such a manner that one or more method steps can be implemented and carried out using a seventh unit.