WO2008138867A2

WO2008138867A2 - Method for managing network components, and a network component

Info

Publication number: WO2008138867A2
Application number: PCT/EP2008/055712
Authority: WO
Inventors: Ingo HÜTTER; Michael Weber
Original assignee: Thomson Licensing
Priority date: 2007-05-09
Filing date: 2008-05-08
Publication date: 2008-11-20
Also published as: WO2008138867A3; DE102007021647A1

Abstract

The invention refers to a method for managing network components (G1...G7) in a network and to a network component, wherein logical connections (1...7) are used to form a logical structure (10) between the network components (G1...G7); the fact that a network component that was disconnected from the network after the logical structure (10) has been formed is missing is detected by at least one detecting one of the network components (G1...G7) that was directly connected to the disconnected network component (G1...G7) in the logical structure (10); the at least one detecting network component (G1...G7) generates at least one message informing that one of the network components (G1...G7) has been disconnected; and all network components (G1...G7) remaining in the network are informed through the at least one generated message that one of the network components (G1...G7) has been disconnected.

Description

METHOD FOR MANAGING NETWORK COMPONENTS, AND A NETWORK

COMPONENT

The invention refers to a method for managing network components and to a network component.

Prior art

Networks, in particular local networks, can occasionally comprise a great number of network components communicating with each other via physical connections, for example via cable connections and/or wireless connections. As an important requirement for a smooth flow of communication in such a network, each of the network components must have specific information about the other network components. In particular, the network components must have knowledge of which other network components are connected to the network.

Time and again during operation of a network, new network components are connected and already connected network components can be disconnected from the network. For that reason, it is very important for the operation of networks and the management of network components that new network components are found and removal of a network component is detected. The pertinent technology is termed discovery wherein methods for discovering new network components and network components that do no longer exist are called discovery methods.

There is a number of such discovery methods, for example discovery methods within the scope of "Universal Plug and Play'' (UPnP), "Multicast Domain Name System'' (MDNS), or "Service Location Protocol" (RFC 2608) . As a matter of principle, the known discovery methods comprise two different embodiments. In a first embodiment, the network components are subdivided in two different classes. A first class comprises the network components to be found by means of the discovery method and a second class comprises the network components that are to find other network components. Where the UPnP protocol is concerned, a subdivision in "control points" and other "devices" is an example thereof. In a second embodiment, all network components are equal, with the result that all of the network components can find all of the other network components in the network.

The known methods are reliable and fast with regard to the discovery of new network components. If, however, network components suddenly disappear from the network without having the possibility of giving prior notice thereof, for example because an Ethernet plug has been pulled off, other network components in the network often take a relatively long time to detect this fact.

In traditional methods, which are provided to detect a separation of network components from a network, each network component emits data packets, which are called discovery data packets, to all other network components of the network at regular intervals. Subsequently, the responses of the other network components to the emitted data packets are waited for. A missing response packet indicates that the associated network component is no longer connected. Invention

It is an object of the invention to provide a method for managing network components and to provide a network component, wherein the detection whether one or more network components have been removed from a network can be achieved quickly and reliably and with a minimum of required computer power. Moreover, it was an object to minimize a network load caused by additional data exchange .

This problem is solved by the invention by means of a method for managing network components according to independent Claim 1 and by means of a network component according to independent Claim 11.

According to the invention, a method is provided for managing network components in a network, wherein logical connections between the network components are used to form a logical structure; the fact that a network component that was disconnected from the network after the logical structure has been formed is missing is detected by at least one detecting one of the network components that was directly connected to the disconnected network component in the logical structure; the at least one detecting network component generates at least one message informing that one of the network components has been disconnected; and all network components remaining in the network are informed through the at least one generated message that one of the network components has been disconnected. According to another aspect of the invention, a network component is provided, comprising connection units configured to form logical connections to other network components; a control unit connected to the connection units and configured to check the logical connections to the other network components; an information unit connected to the connection units and configured to generate a message informing that one of the network components is missing and to transmit the generated message via one of the logical connections; and a transfer unit connected to the connection units and configured to receive a message informing that one of the network components is missing via one of the logical connections and to transmit said message via another one of the logical connections.

The invention comprises the idea of establishing a logical structure in the network at a given point in time, said logical structure including all network components. In this logical structure, there are individual network components which only monitor the existence of neighboring network components. If one of the network components is then removed from the network at a later point, the logical structure consisting of logical connections helps to detect that the removed network component is missing and to transfer this knowledge to all of the remaining network components. Preferrably, this knowledge is transferred to all of the remaining network components via the logical connections of the logical structure. Herein, the disconnection, removal or detachment of the network component from the network is an expression for the disconnection or separation of physical connections of the network component to other network components of the network. Herein, the fact that the removed or disconnected network component is missing can either be detected by preferrably exactly one network component or by a plurality of network components which, in the logical structure, were directly connected to the disconnected network component via an appropriate logical connection. In the event of the missing network component being detected by a plurality of network components, the detecting network components can each generate a message informing that a network component has been disconnected from the network, thus informing the remaining network components through these messages.

Herein, the logical structure does not have to orientate itself by the actual physical connections between the network components of the network. For example, a direct logical connection can be established even between network components which are connected to each other via one or more other network components rather than directly via physical connections. For example, the logical connections can comprise TCP socket connections (TCP - "Transmission Control Protocol") . Moreover, the logical connections can be formed by means of an exchange of data packets according to the UDP ( "User Datagram Protocol") network protocol. Herein, the logical structure may comprise a topography which is optimized for detecting the removal of a network component and for transmitting the message thereabout to the remaining network components .

An advantageous embodiment of the invention provides that the message generated by the detecting network component is transmitted to all of the remaining network components via a plurality or all of the logical connections of the logical structure. The logical structure, therefore, provides an effective means for distributing the information that one network component has been disconnected from the network as quickly as possible throughout the network.

An advantageous further development of the invention provides that the fact that the network component disconnected from the network is missing is detected by the detecting one of the network components by checking one of the logical connections of the logical structure. This check is, preferrably, carried out in a proactive manner. For example, the check of the logical connection can be achieved by sending test packets via the logical connection at certain, for example regular, time intervals .

A preferred embodiment of the invention provides that the logical structure is formed by the network components in a self-organizing manner by establishing the logical connections between the network components. This is to advantage in that it does not require any managing unit which is superordinate to the network.

An appropriate further development of the invention provides that, when forming the logical structure in a self-organizing manner, the network components each use a search function to discover an accessible logical port of another one of the network components and, thereafter, establish the respective logical connection between themselves and the other one of the network components via the discovered accessible logical port. Such logical ports are, for example, TCP sockets. The search can be carried out by means of a traditional group discovery method, which is also called a multicast discovery method. An appropriate group call request is answered only by those network components that have an unassigned logical port at the time of the request. If the searching network component itself has an unassigned or at least accessible logical port, then this port is ignored during the search.

An advantageous embodiment of the invention provides that the message generated by the detecting network component comprises information about the disconnected network component. For example, the message may comprise information about which of the network components has been disconnected from the network, for example by describing its original physical position or its position in the network. Moreover, the message can, for example, provide information about the function of the disconnected network component. This is to advantage in that the network components remaining in the network can prepare themselves for the fact that this function is no longer provided in the network.

If the message about a network component having been disconnected does not provide any further information about the disconnected network component, it is, as an alternative, possible to re-explore the entire network or a partial network, for example by accessing known discovery methods in order to fathom the change in the network topology, resulting from the disconnection of the disconnected network component. Preferrably, a further development of the invention provides that the other network component retrieves and transfers the message generated by the detecting network component. On the one hand, this can be achieved by the message being transferred directly, for example as a message packet. As an alternative, the message can also be evaluated in order that a new message packet can subsequently be formed, which will then be transferred.

A preferred embodiment of the invention provides that the logical structure is formed without any ring topologies. This is to ensure that a closed logical structure is prevented from forming, with the result that at least one logical port of a network component always continues to be available so that, when an additional network component is introduced in the network, there is the possibility to connect the additional network component to the existing logical structure via a logical connection .

An advantageous further development of the invention provides that the logical structure is formed with a linear unbranched topology. Herein, each network component is in communication with no more than two other network components via logical connections, so that the logical structure is formed as a "string" of sorts, which comprises two terminal components. If a network component is now removed from the network, this string is broken up into two partial strings, and the information about the string having been broken up can be transferred very quickly to the remaining network components via the remaining logical connections in the two partial strings. As a result, the data traffic required for applying the method and / or the time required for informing all network components is/are minimized. This relationship will be illustrated below by studying a local network comprising a number n of network components. According to the traditional known methods, a check with regard to whether a network component has left the network is carried out in that each network component emits discovery packets to all other network components and, owing to the incoming response packets returned from the network components arranged in the network, determines whether one of the network components having previously responded does not respond any longer. In this check, each network component sends discovery packets to (n-1) other network components so that, together with the response packets, 2 • n • (n - l)= 2n² - 2n packets have to be transmitted per time unit, wherein a time unit is the time interval between two check cycles.

If, however, use is made of the method described herein and n network components form a linear logical structure, then only 2»(n-i) packets are exchanged per time unit. That means that, while growing quadratically with increasing number n of network components according to the known method, the data traffic grows only linearly according to the method described herein. As a result, either the data traffic required with a defined response time, i.e. with a defined time interval between check cycles, is clearly reduced or, if the data traffic capacity is defined (for example for technical reasons) , the time intervals between the check cycles can be shortened in order that a network component leaving the network is detected clearly faster, that is faster by a factor n.

If, for example, the network comprises 100 network components exchanging information in the traditional manner at an interval of 5 minutes, then, as a result, it takes 5 minutes on average until a network component is detected to have left the network. According to the method described herein, however, this takes only 3 seconds, starting from the assumption that the data traffic is uniform.

An appropriate further development of the invention provides that the logical connections between the network components are formed as packet-based connections. As an alternative thereto, the logical connections can be formed on a line transmission basis.

Drawing :

Below, the invention will be illustrated in more detail by means of exemplary embodiments with reference being made to figures of a drawing, in which:

Fig. 1 is a schematic diagram of the structure of a network component with logical ports;

Fig. 2 is a schematic diagram of a logical structure comprising a plurality of network components; and Fig. 3 is a flow diagram of a method for generating a logical structure in a network.

Exemplary embodiments:

Fig. 1 is a schematic diagram of the structure of a network component G with logical ports P, A. Using the logical ports P, A, the network component G is able to enter into logical connections 1 to other network components (not illustrated) in order to establish a logical structure. The logical ports P, A of the network component G comprise a passive port P serving to accept logical connections from other network components. That means that a logical connection running from the other network component to the network component G via the passive port P is initiated by the other network component. Furthermore, the logical ports P, A of the network component G comprise an active port A which can be used to establish a logical connection 1 to a passive port of another network component (not illustrated) . To achieve this, a search process is run through the network component G in order to find accessible passive ports of other network components in a network.

Examples of a logical connection 1 between the network component G and another network component are TCP socket connections which are used to exchange data at regular intervals. Furthermore, the logical connection 1 can be formed by exchanging UDP packets between the network component G and the other network component.

If the network component G is added to a network of other network components by connecting the network component G to one or more of the network components, the search process is run through the network component G in order to find those network components in the network which comprise accessible passive ports. The network component G will then establish a logical connection 1 between its active port A and the passive port of the network component found, in order form a logical structure. Such search processes are known from traditional discovery methods .

Fig. 2 is a schematic diagram of a logical structure 10 comprising network components G1...G7 which are connected to each other by means of logical connections 1...7. As shown in Fig. 2, a logical connection 2 is, for example, established between the network component G2 and the network component G3 by forming the logical connection 2 between the active port A of the network component G2 and the passive port P of the network component G3. The logical connections 1...7 are used to exchange information for monitoring the connections 1...7 between the network components G1...G7. Furthermore, the logical connections 1...7 can be used to exchange information which is used to check whether the logical structure 10 formed by means of the logical connections 1...7 comprises a ring topology. By avoiding a ring topology, it is ensured that network components G1...G7 with accessible passive ports P are always available in the logical structure 10, wherein new additional network components having been introduced in the network can use their active ports to establish logical connections.

In the logical structure 10 shown in Fig. 2, the network components Gl and G6 comprise passive ports P which are accessible at any rate. It is true that a passive port P can establish logical connections 1...7 to a plurality of active ports A, as is the case with the passive port P of the network component G5 in the logical structure 10 shown in Fig. 2. The network component G5 is connected to the network components G4 and G7 by means of the logical connections 4 and 7, with the result that the logical structure 10 comprises a branched topology. If, however, a sufficient number of such branches is available in the logical structure 10 and the logical structure 10, furthermore, comprises a ring topology or substructures with ring topologies, then this may result in no other accessible passive ports being available in the logical structure 10. It is, therefore, necessary to avoid ring topologies. Preferrably, branched topologies as used in the logical structure 10 shown in Fig. 2 are avoided as well, so that the logical structure 10 will then only comprise a linear topology.

A logical structure 10 without any ring topology is to further advantage in that a message transferred along the logical connections 1...7 by the network components G1...G7 will, at some point, reach the end of the logical structure 10. In a ring topology, additional measures would have to be taken to prevent the message from running repeatedly through the logical structure 10.

A possibility to check whether there is a ring topology will be described below. A network component G1...G7 having established a logical connection 1...7 to another network component G1...G7 by means of its active port A emits a data packet via the logical connection 1...7. The emitted data packet is transferred by means of the other network component G1...G7. If the data packet arrives at a network component G1...G7 the active port A of which is not connected, the data packet is marked appropriately and returned. In this manner, the network component G1...G7 having emitted the data packet detects that at least one substructure of the logical structure 10 in which it is arranged comprises a terminal point so that the substructure does not have any ring topology. If, however, the emitted data packet arrives at the passive port P of the network component G1...G7 which emitted it, the network component G1...G7 detects that it has formed a logical structure 10 or a substructure (not illustrated) with a ring topology by means of the logical connection 1...7 established via its active port A. It can then disconnect the logical connection 1...7 to avoid the ring topology and can, subsequently, search for another network component G1...G7 with an accessible, for example unused and, therefore, unassigned passive port P.

After the logical structure 10 has been formed by means of the logical connections 1...7, the logical structure 10 can be used to detect a disconnection of a network component G1...G7 from the network and to inform the remaining network components G1...G7 thereof via the logical connections 1...7. The procedure required to this end will be illustrated below by means of an example assuming that the network component G3 has been disconnected from the network, for example by simply pulling off the network plug or by switching off the network component G3.

The network components G1...G7 check each of the particular connections 1...7 running to them at regular intervals. Therefore, the logical connections 2 and 3 are regularly checked by the network components G2 and G4, for example by exchanging data packets between the network components G2 and G3 on the one hand and between the network components G3 and G4 on the other hand. For that reason, the non-existence of the network component G3 will be detected by the network components G2 and G4 within a short time, because the connections 2 and 3 will then not be functioning any longer.

Thereafter, the network components G2 and G4 each prepare messages comprising information about the fact that the network component G3 is missing. This information may comprise further additional information about the network component G3, for example information about the functions executed by the network component G3 in the network. Subsequently, the messages are sent to the remaining network components (network components Gl, G5, G6, and G7 in Fig. 2) via the remaining connections (connections 1, 3, 4, 6, and 7 in Fig. 2) . Preferrably, a discovery method is then re-applied in order to determine whether a plurality of coherent network components G1...G7 has been separated from the network.

Since the network component G3 is no longer available, there is now a logical structure 10 which is divided in two partial structures (not illustrated) , that are a partial structure with the network components Gl and G2 and a further partial structure with the network components G4, G5, G6, and G7. Since there is no longer any logical connection 2, the active port A of the network component G2 is unassigned, with the result that the network component G2 must run through the search process in order to find another network component with an accessible passive port, for example the network component G4 with the passive port P which is now open because the logical connection 3 to the network component G3 has been separated. As an alternative, the search process can also be used to find the network component G6 with the unassigned passive port P. It is true that the network component Gl also comprises an unassigned passive port P, but a connection between the active port A of the network component G2 and the passive port P of the network component Gl would result in a logical structure with a ring topology, which is avoided according to the method described above.

Fig. 3 is a flow diagram of a method for generating a logical structure 10 in a network with network components G1...G7. The method is applied by a network component G1...G7 in order to form a logical connection 1...7 to another network component G1...G7. After the method shown in Fig. 3 has been applied in all network components G1...G7 of the network either successively or at least simultaneously, a logical structure 10 will be established, which does not comprise any ring topology.

The method is started in a first step 20. Then, it is checked 21 whether the active port A is already connected. If this is the case, the system waits in a step 22 until the active port A is not connected any longer. This can, for example, be achieved by separating a network component G1...G7 from an already formed logical structure 10. If, however, the active port A is not connected, all network components G1...G7 comprising an accessible passive port P are searched for in the network in a next step 23. If this search criterion is limited not only to accessible passive ports P but even to unassigned passive ports P, that means to those passive ports that are not connected at all, then a linear unbranched structure can be formed according to the present method.

Then, it is checked whether at least one network component G1...G7 with an accessible passive port P has been found 24. If this is not the case, a waiting loop with defined duration is run through in a further step 25, in order to subsequently repeat 26 the search. If, however, accessible or unassigned passive ports P have been found, the method runs through a loop 27, 35 in which an attempt is made to establish a logical connection 1...7 to one of the found network components G1...G7 having an unassigned or accessible passive port P. To achieve this, a logical connection between the active port A and the found passive port P is formed for each found passive port P in a connection step 28. In a next step 29, it is checked whether the logical connection has been established successfully. If this is not the case, for example because another network component G1...G7 established a connection to the found passive port P in the meantime, the other found passive ports P are examined in similar manner. If the logical connection 1...7 was established successfully in step 29, it is checked in a further step 30 whether the logical connection 1...7 thus formed is part of a logical structure 10 with ring topology, that means whether this connection is what is called a circular connection. This is, for example, achieved by emitting a data packet as described above. If a ring topology or a ring is, thus, available 31, the connection is separated again in a next step 32.

If this is not the case, the logical connection formed is maintained and a change in topology of the network is waited for 33. A change in topology is, for example, achieved by removing network components G1...G7 from the network or by adding further network components to the network. Such a change in topology, in particular caused by adding further network components to the network, can be detected or examined by means or with the help of traditional discovery methods which are typically based on group call or multicast methods.

Step 31 in which it is determined whether a ring topology or a ring has been formed can, optionally, also be used to determine how many network components G1...G7 participate in the ring, that means to determine the size of what is called the ring length. To achieve this according to the method described above for checking whether there is a ring topology, a counter which is incremented when running through a network component G1...G7 is dispatched together with the emitted data packet. Once the data packet has again arrived at the original network component G1...G7, the counter reading corresponds to the ring length of the ring formed.

The elements of the invention disclosed in the foregoing description, the claims and the drawing can be of import to the implementation of the invention in its different embodiments both individually and in any combinations desired.

Claims

1. A method for managing network components (G1...G7) in a network, wherein logical connections (1...7) are used to form a logical structure (10) between the network components (G1...G7); the fact that a network component disconnected from the network after the logical structure (10) has been formed is missing is detected by at least one detecting one of the network components (G1...G7) that was directly connected to the disconnected network component (G1...G7) in the logical structure (10); at least one message informing about one of the network components (G1...G7) having been disconnected is generated by the at least one detecting network component (G1...G7); and all network components (G1...G7) remaining in the network are informed about one of the network components (G1...G7) having been disconnected due to the at least one generated message .

2. The method according to Claim 1, characterized in that the generated message is transmitted to all of the remaining network components (G1...G7) via a plurality or all of the logical connections (1...7) of the logical structure (10).

3. The method according to Claim 1 or 2, characterized in that the fact that the network component disconnected from the network is missing is detected by the detecting one of the network components (G1...G7) by checking one of the logical connections (1...7) of the logical structure (10).

4. The method according to anyone of the preceding claims, characterized in that the logical structure

(10) is formed by the network components (G1...G7) in a self-organizing manner by establishing the logical connections (1...7) between the network components

(G1...G7) .

5. The method according to Claim 4, characterized in that, when forming the logical structure (10) in a self-organizing manner, the network components (G1...G7) each use a search function to discover an accessible logical port (P) of another one of the network components (G1...G7) and, thereafter, establish the respective logical connection (1...7) between themselves and the other one of the network components (G1...G7) via the discovered accessible logical port (P) .

6. The method according to anyone of the preceding claims, characterized in that the message generated by the detecting network component (G1...G7) comprises information about the disconnected network component (G1...G7) .

7. The method according to anyone of the preceding claims, characterized in that the other network component (G1...G7) retrieves and transfers the message generated by the detecting network component

(G1...G7) .

8. The method according to anyone of the preceding claims, characterized in that the logical structure (10) is formed without any ring topologies.

9. The method according to Claim 8, characterized in that the logical structure (10) is formed with a linear unbranched topology.

10. The method according to anyone of the preceding claims, characterized in that the logical connections (1...7) between the network components (G1...G7) are formed as packet-based connections.

11. A network component comprising connection units configured to form logical connections (1...7) to other network components (G1...G7); a control unit connected to the connection units and configured to check the logical connections (1...7) to the other network components (G1...G7); an information unit connected to the connection units and configured to generate a message informing that one of the network components (G1...G7) is missing and to transmit the generated message via one of the logical connections

(1...7); and a transfer unit connected to the connection units and configured to receive a message informing that one of the network components (G1...G7) is missing via one of the logical connections (1...7) and to transmit said message via another one of the logical connections (1...7).