EP2183889A1

EP2183889A1 - Method of establishing constraints on service stream routing in a virtual local network

Info

Publication number: EP2183889A1
Application number: EP08784735A
Authority: EP
Inventors: Christophe Le Guern
Original assignee: Sagem Communications SAS
Current assignee: Sagemcom Energy and Telecom SAS
Priority date: 2007-07-13
Filing date: 2008-07-11
Publication date: 2010-05-12
Also published as: FR2918831B1; WO2009010243A1; FR2918831A1

Abstract

The present invention relates to a method of establishing constraints on service stream routing in a local or infrastructure network consisting of interconnected bridges, said routing constraints defining those of the internal links between ports of each of said bridges whose establishment is prohibited, one of said bridges being considered to be the network head bridge, said method comprising a step of associating port attributes of bridges of the network, in the course of which a root attribute is associated with the port, which affords the most direct access to said network head bridge, characterized in that it comprises: - a step (200) of obtaining said associations between ports of each bridge and port attributes, and - a step (300) of defining routing constraints relating to each bridge in the course of which the establishment of any internal link between two ports of a bridge is prohibited if one of these two ports is not associated with the root attribute.

Description

Method for establishing service flow routing constraints in a VLAN

The present invention relates to a method of establishing service flow routing constraints in a local network or infrastructure consisting of interconnected bridges.

Local or infrastructure networks are known that route Ethernet-based service flows either between a first user's equipment and a second user's equipment, or between a first user's equipment and the equipment of a first user. several users in the case of point-to-point and point-to-multipoint communications or multipoint-to-multipoint communications, respectively. In the remainder of the description, the term network will be used to designate a physical network that is local or infrastructure.

To overcome the constraints of the physical architecture of a network

(geographic constraints of equipment, addressing constraints, etc.), it is common to determine from this architecture one or more so-called virtual networks. These virtual networks are commonly called VLANs (Virtual Local Area Network) in the case of local networks but this VLAN designation is also conventionally used to designate a virtual network determined from an infrastructure network.

A virtual network defines a logical topology of the physical architecture of a network. For this purpose, the user equipment (subscribers, service providers, professional users, etc.) connected to this network are first grouped either on the basis of physical criteria such as the MAC (Medium Access Control) address. ) specific to each user equipment, either on the port numbers specific to each user equipment, or on the basis of services such as layer 2, 3 or higher of the OSI model, or on the basis of protocols used by each user equipment to communicate.

The user equipment of a network is most often interconnected by equipment called bridges.

Thus, once the user equipment groupings made from the physical architecture of the network, the groups of user equipment obtained are interconnected logically between them by these bridges to form a virtual network. It can be noted that a virtual network can also be created for each equipment group. Virtual networks connected by bridges are then obtained.

Most bridges of a network allow access to services because they are also connected to one or more groups of equipment and then allow the exchange of service flows between user equipment belonging to this or these groups. In addition, one of these bridges, called the headend bridge, is also connected to a service provider's equipment to provide service flows to the user equipment of all groups. Subsequently, we will consider so-called Ethernet bridges with reference to the type of service flow frames they carry. The virtual network considered thereafter will be a layer 2 virtual network.

An Ethernet bridge is a device that has a number of ports (at least two) to which user devices or certain other bridges of the network are connected. In addition, each Ethernet bridge includes a bridge matrix that informs the bridge of the internal link to establish between two (or more) of its ports when a service flow frame is to be transferred from user equipment (or bridge) to another user equipment (or to another bridge). For this, each bridge undergoes a learning phase that allows to know the equipment or bridges that are connected to each of its ports. The transfer matrix then memorizes the relationships between each port and the addresses of the user equipment or bridges that are connected to this port.

A network can only work if the paths between bridges of the Ethernet frames do not form loops, which occurs when an exchanged frame passes twice through the same bridge. Indeed, when broadcast frames will be sent on such a network, each bridge sends them through all its ports. The frames then circulate in loops and are multiplied each time by a bridge. The frames have no life, so they can turn indefinitely in one of the loops of the VLAN. This is commonly known as the "broadcast storm".

In order to prevent such situations from occurring, it is necessary to define a virtual network whose logical topology is arborescent, that is to say which has no loop. For this purpose, it is customary to use a spanning tree protocol (STP) defined by the ISO / IEC 15802-3: 1998 standard (IEEE Std 802.1D-1998) or one of its extensions such as the protocol Rapid Spanning Tree Protocol (RSTP) defined by Annex 802. Iw of the IEEE Std 802.1D-2004 standard or the Multiple Spanning Tree Protocol (MSTP) defined by the IEEE Std 802.1s standard integrated in the 802.1 standard. .

We will limit ourselves to describing the principle of the STP protocol in the case of a simple network in order to show how a loop of a network can be suppressed. We are talking about opening a loop. Nevertheless, the STP protocol as well as its extensions make it possible to open much more complex network loops.

The STP protocol allows the regular exchange between the bridges of the network of protocol units called BPDUs (Bridge Protocol Data Unit in English) in order to obtain a tree topology of the virtual network or so that a new topology of this virtual network can be determined when an event occurs on the network (connection of a new equipment, rupture of a segment between two bridges, failure of a bridge, ...).

STP is a Layer 2 protocol, that is, it operates directly at the Ethernet bridges of the network. This protocol is said convergent because from a finite number of exchanges of BPDU units, the role of each bridge network is defined to determine the paths taken by the Ethernet frames.

For this, the STP protocol is broken down into three successive steps which are respectively an election stage of the root bridge (root bridge in English), a step associating the port attributes and a step of assigning the port states according to the port attribute associations.

During the root bridge election stage, the network bridges agree to elect one of them as a root bridge. For this, each bridge B ₁ of the network is associated with a unique identifier IDB (Bridge Identifier in English).

Each bridge B, also keeps in memory a parameter IDR ₁ (Root Identifier in English) likely to take the value of a bridge identifier. The identifier IDB ₁ of a bridge has two parts: firstly, the priority (expressed on 2 bytes) and, on the other hand, the MAC address of the bridge (expressed on 6 bytes). The priority of a bridge is IEEE 802. The default ID is equal for all bridges. However, as we will see later, the administrator may have an interest in defining different priority values according to the bridges.

At the beginning of the root bridge election step, the IRD parameter ₁ of each bridge

B ₁ is equal to the identifier IDB ₁ of the bridge B ₁ which stores it. Each bridge B ₁ of the network then considers that it is the root bridge of the virtual network. Then begins sending a BPDU by one of the bridges of the virtual network, for example by the bridge B ₁ , to the other bridges to which it is connected.

Suppose a B _{1 + I} bridge receives this BPDU. The bridge B ₁₊ ] then compares the IDB bridge identifier ₁ contained in the received BPDU unit with its parameter IDR ₁₊ ] and two cases can then occur.

In the case where the bridge identifier IDB ₁ received by the bridge B _{1 + I} has a value lower than that of the parameter IDR ₁₊ I that it stores, the value of its parameter IDR _{1 + I} is replaced by the IDB bridge ID ₁ received. The bridge B _{1+ 1} then considers that the root bridge is the bridge B ₁ . In the case where this bridge identifier IDB ₁ has a value greater than that of its parameter IDR _{1 + I} , the bridge B, ₊ i does not make any substitutions and continues to consider that the root bridge is the one designated by the value of its IDR _{1 + I} parameter.

By exchanging their IDR parameter values ₁ via BPDUs, the bridges B ₁ of the virtual network all end up having the same value of parameter IDR, and consequently to elect by common agreement a single bridge to rank root bridge of the virtual network.

The step of association of the attributes of port begins following the stage of election of the root bridge. The purpose of this step is that a so-called port attribute is associated with each port of each bridge of the virtual network. The attribute associated with a port defines the role played through this port in the determination of the paths of the Ethernet frames through the virtual network.

When a port is able to route Ethernet frames arriving from the root bridge to the user equipment of the group to which it is connected and to route the frames arriving from these user equipments to the root bridge, the associated attribute to this port gives this port the so-called transfer status. According to the IEEE Standard Std 802. ID-2004, the port attributes that give this port the transfer status are the attributes 'designated' {designated in English) and 'root' (root). These two terms are to be considered later within the meaning of the IEEE Standard Std 802.1D-2004. When a port can not route Ethernet frames arriving from the root bridge to the user equipment of the group to which it is connected and neither route the frames arriving from these user devices to the root bridge, the attribute associated with this port gives this port the so-called blocking state. According to the IEEE Std 802.1D-2004 standard, the port attributes that give this port the blocking state are the 'alternate' (also known as 'backup') attributes.

During the step of association of the attributes of port, the protocol STP associates with each port of the various bridges the attribute is designated (D), is root (R), or alternates (A) which will make it possible to define if this port is in the state of either transfer or blocking. Note that the root attribute is associated with a port of each bridge other than the root bridge. For this, the port of a bridge is elected to the rank of root port when it is in connection with the root bridge and allows that the exchange of frames between this bridge and the root bridge is the fastest among the possible exchanges between the other ports of this bridge and the root bridge. In other words, the root port of a bridge is the one that provides the most 'direct' access of this bridge to the root bridge. In addition, the designated attribute is associated with at least one port of each bridge. Finally, the root bridge ports are all associated with the 'designated' attribute. We can refer to the IEEE Std standard. 802. ID for more details.

Once the association of the designated attribute or root is performed, the ports of each bridge that are not associated with either the root attribute or the designated attribute are associated with the alternate attribute.

The step of assigning the port states follows the step of associating the port attributes. In this step, each port associated with the designated attribute or attribute root is considered in the transfer state and each port associated with the alternate attribute is considered in the blocking state.

Fig. 1 is a diagram of an example of a NET virtual network. The virtual network NET consists of five groups of equipment referenced G ₁ (i = 1 to 5) respectively connected to the five Ethernet bridges B, (i = 1 to 5) that comprises the network from which the virtual network NET has been got.

Each Ethernet bridge B ₁ , according to this example comprises three ports P, _j (j = 1 to 3) and a transfer matrix referenced Mt ₁ which defines the internal links I _{1 k} (k = 1 to 3) between the three ports P _(J of each bridge B _1. It may be noted that to facilitate the disclosure of the invention, an internal link I ₁ ^ is here considered as linking a port of a bridge to only one other of the ports of this bridge. , an internal link can also be considered as linking a port to several other ports of the same bridge without the invention being modified, each arrow inside the transfer matrix Mt ₁ means The link I ₁ can be established in Fig. 1. Only the links of the bridge Bi are referenced li _{, k} (k = 1 to 3).

The present invention is not limited to three-port Ethernet bridges because the only constraint imposed on these bridges is that the number of their ports is greater than or equal to 2.

Each equipment group comprises a certain number of user equipment of which only two are shown in FIG. 1: the user equipment EUi situated in the group Gi and the user equipment UE ₂ located in the group G ₂ . In addition, the equipment group G ₅ comprises an equipment E of an operator or service provider. This equipment is intended to provide service flow frames for all user equipment of all equipment groups.

The topology of the virtual network NET is a ring topology that is to say that each bridge of the virtual network is connected to two other of its bridges. This logical topology is used for example when the physical architecture of this network is of the sonet / SDH type. This example of topology has been chosen to facilitate the explanations that follow. However, the present invention is not limited to this type of logical topology but can be applied to virtual networks whose topology is more complex than that of FIG. 1, for example when each bridge is connected to more than two other bridges. In addition, the present invention is not related to the type of the physical architecture because the present invention is in layer 2 of the OSI model while the physical architecture is in layer 1 of this model.

Fig. 2 represents an example of the NET virtual network following a convergence of the STP protocol. According to the example given by this figure, the ports Pi ₂ , P _{2 2} , P ₃ , ₂ and P ₄₃ are associated with the root attribute R and the ports Pi _i3 , P _2j3 , P _3j3 , P _{5i 2} and P _{5 3} are associated with the designated attribute D. Finally, the port P ₄₂ which is associated with the alternate attribute A is in the blocking state thus logically opening the loop formed by the ring topology of this virtual network ( the segment 'blocked' between the ports P _{4 2} and P _{3; 3} is in dotted line in Fig. 2). The port of each bridge connecting a user equipment group to this bridge is associated with the designated attribute D so that this port is in the transfer state.

It can be noted that the STP protocol continuously monitors the evolution of the logical topology that it has just determined. We usually talk about active topology. For this purpose, each port in the transfer state of a bridge B ₁ regularly sends a BPDU to the port of another bridge B _{j to} which it is connected.

This other port can be in the transfer state but also in the blocking state. Thus, since a port of the bridge B _j which is likely to receive at regular intervals these BPDU units no longer receive these units, the STP protocol then considers that an event has occurred on the segment connecting these two ports. and restart a convergence to obtain a new logical topology of the virtual network without taking into account this segment.

The STP protocol determines that one of the bridges of a virtual network is elected root bridge. It is possible to force the STP protocol to always elect the same bridge as a root bridge. This is useful, in particular, for this root bridge to be the bridgehead of the network. For this, the STP protocol electing the root bridge from the identifiers associated with these bridges and these identifiers being defined, among others from a priority value, the administrator can force this protocol to always elect the root bridge as being the head bridge by assigning the lowest possible priority value to this head bridge.

The NET network administrator, when configuring this network, determines particular paths taken by service flow Ethernet frames by allowing or preventing these frames from being routed through the internal links of the different bridges of the NET virtual network. . Fixed administrator thus constraints called flow of service flow. Once the routing constraints have been defined, the administrator manually configures the transfer matrix Mt ₁ of each bridge B, according to the routing constraints chosen, and this, by means of a network manager. According to the example of FIG. 2, each internal link l, _k of each Ethernet bridge

B, is allowed, that is to say that no particular routing constraint is imposed on the transfer matrix Mt ₁ of each bridge B ₁ . The UEi user equipment can then directly exchange Ethernet frames with the user equipment UE ₂ via the bridges Bi and B ₂ without their exchanges passing through the headend bridge B ₅ on which a For example, a service provider may want to connect a device C for controlling flow exchanges between users. This device C is shown in FIG. 2 as being part of the equipment E.

To overcome this problem of direct exchange of user equipment flows, an exchange that is not desired by operators and other service providers who wish to protect any fraudulent use of their service flows, the network administrator defines service flow constraints that force each exchange of flows to pass necessarily through the headend bridge.

Fig. 3 represents a diagram of the virtual network NET when such constraints have been imposed during the configuration of this network. According to these routing constraints, the establishment of the links li _j2 and I _{1 3} of the bridge Bi is authorized while the establishment of the internal link li i of this bridge Bi (connection between the ports Pi _, i and Pj ₃ ) forbidden. This routing constraint is schematically illustrated in FIG. 3 by removing the arrow between the ports Pi ₌ i and P _{I 3} . It may be noted that according to the example of FIG. 3, the bridge transfer matrices B, other than the root bridge B ₅ , were also constrained by a constraint similar to that imposed on the transfer matrix Mtj. Thus, when the UEi and UE ₂ user equipments exchange Ethernet frames, these frames routinely pass through the headend bridge B ₅ and the operator can then control the flow exchanges between these two user equipments from of the device C.

Consider now the case where an event occurs on the network when it corresponds to the virtual network NET as configured according to the example of FIG. 3. This event is for example the breaking of the segment between the bridges Bi and B ₅ , illustrated in FIG. 3 by the cross referenced X. The exchanges in progress between the User equipment EU ₁ and EU ₂ , among others, are then interrupted or made impossible because of the routing constraints. These constraints prevent Ethernet frames from being exchanged between the bridges Bi and B ₂ and furthermore the routing constraint is no longer respected. The bridge B ₅ is now inaccessible via the segment connecting the bridges Bj and B ₅ .

The STP protocol, as we have seen previously, then automatically initiates a new convergence as soon as this event is detected in order to determine a new logical topology of the NET virtual network, that is to say new associations of the port attributes. and determining the states of the different ports of the bridges of the virtual network NET taking into account the break of the segment connecting the bridges Bi and B ₅ .

Fig. 4 represents a diagram of the result of this new convergence of the STP protocol. The ports Pi ₃ , P ₂ 3, P _3j3 and P ₄₃ are now associated with the root attribute R and the ports P _2j2 , P _3i2 and P _{4; 2} are now associated with the designated attribute D. Finally, the port Pi ₂ now associated with the alternate attribute A is in the blocking state.

However, the NET virtual network still does not work despite this new convergence of the STP protocol because the user equipments of the Gi G ₂ and G _{3 groups} no longer have access to the network because of the definition of the routing constraints. It is necessary for the administrator to manually re-define the routing constraints via the network manager as quickly as possible so that users do not have a break in service access that is prolonged.

This update of the routing constraints is generally tedious for the administrator who will have to intervene on each bridge of the network. In practice, the network administrator most often prefers not to modify the routing constraints but rather to repair the event as quickly as possible. Such a strategy implies a risk of loss of quality of service important for users in the event that such a repair is extended over time.

The aim of the present invention is to avoid that the introduction of routing constraints in a network makes it tedious for its administrator, the reconfiguration of a virtual network, determined from this network, and causes a loss quality of service when an event occurs on the network.

For this purpose, the present invention relates to a method of establishing service flow routing constraints in a network of bridges. interconnected between them, said routing constraints defining those internal links between ports of each of said decks whose establishment is prohibited, one of said bridges being considered to be the head-end bridge, said method comprising a step of association of bridge port attributes of the network in which is associated with the port, which provides the most direct access to said headend bridge, a root attribute. The method is characterized in that it comprises:

a step of obtaining said associations between ports of each bridge and port attributes, and

a step of defining the routing constraints relating to each bridge during which the establishment of any internal link between two ports of a bridge is forbidden if one of these two ports is not associated with the root attribute.

Advantageously, said port attribute association step of a bridge being executed as soon as an event occurs on said network, said step of obtaining said associations between ports of each bridge and port attributes is preceded by a step detecting the updating of the attributes associated with the ports of at least one of said bridges.

Thus, the method according to the present invention can also be used when obtaining a new logical topology of a virtual network determined from said network as soon as an event occurs on this network. According to one embodiment, said step of defining the routing constraints comprises a step of associating the root attribute associated with a port of a bridge with a so-called leading attribute and with each other attribute of a port. a bridge to a subscriber attribute, the establishment of an internal link between two ports of a bridge is then prohibited if one of these two ports is not associated with the head attribute. Advantageously, each bridge being associated with an identifier formed inter alia of a priority value, the bridge which is considered to be the headend bridge is the bridge associated with the identifier whose priority value is the lowest among the priority values of the identifiers associated with the other bridges.

Thus, by assigning the lowest priority value to the bridge of said network which is connected to a service operator's equipment, one and the same bridge will always be considered as the head bridge and the operator can thus connect to this bridge head equipment that will allow it to deliver and / or control the flow of service that pass through the network even if these flows pass between users. According to one embodiment of the method, said routing constraints are defined from a virtual network determined from said network, a logical topology of said virtual network is then obtained following a convergence of an STP type protocol. According to one of its material aspects, the present invention also relates to a device for setting service flow routing constraints intended to be associated with an Ethernet bridge of said network subjected to routing constraints.

In another of its aspects, the present invention relates to a computer program for establishing service flow routing constraints of an Ethernet bridge of a network subject to said routing constraints. The said computer program implements the steps of the above method of establishing.

The characteristics of the invention mentioned above, as well as others, will appear more clearly on reading the following description of an exemplary embodiment, said description being made in connection with the attached drawings, among which: FIG. . 1 which represents a diagram of an example of a virtual network, FIG. 2 which represents an example of the virtual network obtained from the network of FIG. 1 following a convergence of the STP protocol, FIG. 3 which represents a diagram of the virtual local network of FIG. 2 when routing constraints were imposed during the configuration of this network, FIG. 4 which represents a diagram of the result of a new convergence of the STP protocol following an event occurring on the network of FIG. 3 according to the state of the art, FIG. 5 which represents a diagram of the result of a new convergence of the STP protocol following an event occurring on the network of FIG. 3 according to the present invention, FIG. 6 which is a flowchart of the steps of the service flow routing constraint method according to the present invention, and FIG. 7 which shows a block diagram of a service flow routing setting apparatus according to the present invention.

The proposed invention is to define service flow routing constraints in a network, based on the port attributes associated with the service ports. each bridge of this network. Thus, as soon as a new port attribute is associated with a port of a bridge, especially as soon as an event occurs in the network and the port attributes have been updated automatically, the routing constraints are then re-defined automatically according to rules for establishing internal links between the ports of each of these bridges.

Fig. 6 is a flow diagram of the steps of the service flow routing constraint method according to the present invention.

This method is intended to be executed by each bridge of a network as soon as a virtual network has been determined and at each port of each bridge has been assigned either a root attribute R, designated D, or alternating A.

The method starts with a step 200 of obtaining the associations between attributes and ports of a bridge B ₁ . Following the example of FIG. 2, during step 200, three associations are obtained for each bridge B ₁ , in particular the associations between the port Pij and the designated attribute D, the association between the port Pi ₃ and the designated attribute D and the association between the Pi ₂ port and the R root attribute.

Step 200 is followed by a step 300 of definition of the routing constraints relating to each bridge. At each bridge B ₁ of the network can be associated with one or more routing constraints.

During the step 300 of definition of the routing constraints relating to a bridge B ₁ , the establishment of any internal link l ₁ is prohibited _; k between two ports of a bridge B, if one of these two ports is not associated with the root attribute R.

Take the example of the bridge B] defined according to the configuration example of the virtual network NET given in FIG. 2. Three links I] ₁ , li _{; 2} and li _i3 can potentially be established according to the frames to be transferred and the contents of the transfer matrix Mti as explained in the introductory part. During the step 300, the establishment of the link Ii j defined between the ports Pi i and Pi ₃ is prohibited according to the rule previously stated because neither the port Pi _, i _, nor the Pj ₃ are associated with the root attribute

A. On the other hand, the establishment of the link I] ₂ which is defined between the ports Pi, i and Pi _{; 2} is authorized, according to the rule previously stated, because the port Pi ₂ is associated with the root attribute R. It is the same for the link I] ₃ .

The method of establishing routing constraints described above can be used during the configuration of a virtual network, thus avoiding an administrator having to manually configure the transfer matrices of the various bridges of this network. However, the method can also be used when the configuration of a new logical topology of the virtual network, obtained as soon as an event occurs on this network.

Thus, it is advantageous for step 200 to be preceded by a step 100 of detecting the updating of the attributes associated with the ports of at least one of said bridges.

Fig. 5 illustrates the result of the execution of the routing constraint setting method according to the example of FIG. 4. In particular, concerning the bridge Bi, the link li i which was prohibited according to FIG. 4 is now allowed and the link I] ₂ which was previously allowed is now prohibited according to FIG. 5. According to an embodiment of step 300, said step 300 comprises a step of associating the root attribute R (associated with a port of a bridge) with a so-called head attribute H {Head in English and each other attribute of a port of a bridge to a subscriber attribute S (Subscήber in English), the establishment of an internal link between two ports of a bridge being then prohibited if one of these two ports are not associated with the H head attribute.

Advantageously, each bridge being associated with an identifier formed inter alia of a priority value, the bridge which is considered to be the headend bridge is the bridge associated with the identifier whose priority value is the lowest among the priority values of the identifiers associated with the other bridges. According to one embodiment of the invention, the steps 200 and 300 being implemented by a computer program, said detection step 100 is implemented by software interrupt.

Fig. 7 is a block diagram of a service flow routing set-up device D according to the present invention. The device D is intended to be associated with an Ethernet bridge B ₁ of a network subject to said routing constraints. According to the embodiment shown in FIG. 7, the device D is integrated in the bridge B ₁ but, according to another embodiment, the device D can only be associated with the bridge B ₁ .

The bridge B ₁ comprises a communication bus B to which are connected means PROC for processing digital data, a nonvolatile memory ₁₅ ROMi a RAM ,, P ports, _j associated to COM means for receiving and transmit service flow frames, MTT means ₁ for transferring an incoming frame on a port P _g to one or more other ports P _{1; k} and MAP means for associating an attribute with each of its ports and, in particular at the port that provides the most direct access to said head bridge the root attribute R. The MAP means implement the STP protocol. The bridge B ₁ also includes a transfer matrix Mti which stores, on the one hand, the data defining the relationships between equipment (and / or bridges) and ports of this bridge, relationships that it has learned during a phase and on the other hand, the associations between ports of the bridge B, and port attributes determined by the MAP means ₁ .

The device D comprises means MOA for obtaining said associations between ports of the bridge B ₁ and port attributes, stored in the transfer matrix Mti and means MDA for defining transport constraints between the ports of the bridge B ₁ as a function of the port attributes associated with each bridge port B ₁ .

According to the present invention, the means MDA prohibit the establishment of a link between two ports of the bridge B ₁ if one of these two ports is not associated with the root attribute R.

According to another embodiment of the means MDA, each root attribute R associated with a port of the bridge B ₁ is associated with the head attribute H and each other attribute of a port of the bridge B ₁ is associated with the attribute d Subscriber S. The establishment of an internal link between two ports of a bridge is then prohibited if one of these two ports is not associated with the attribute of head H.

According to one embodiment, a PRG1 program being used to implement the MAP means, as well as the data accesses of the transfer matrix Mti, a program PRG2 then implements the means MOA and MDA.

The non-volatile memory ROM ₁ then stores the PRG1 and PRG2 programs and digital data allow, among other things, their implementation. More generally, the ROM ₁ is readable by the processing means PROC ₁ , integrated or not to the bridge B ₁ and can be removable.

The memory ROM thus stores the associations between port attributes and ports of the bridge B ₁ and the data of the transfer matrix Mt ₁ obtained, on the one hand, during a prior learning phase and on the other hand the data defining the routing constraints imposed on this transfer matrix. When powering up the bridge B ₁ , the programs according to the present invention are transferred into the RAM RAM ₁ which then contains the executable code of the invention as well as the data necessary for the implementation of the invention. Thus, as soon as a port attribute is newly associated with a port by the PRG1 program, this program is interrupted and the program PRG2 is launched in order to redefine the routing constraints relating to this port. The program PRG1 then resumes its execution as soon as the program PRG2 ends.

Claims

1) Method for establishing service flow routing constraints in a local network or infrastructure consisting of interconnected bridges, said routing constraints defining those of the internal links between ports of each of said bridges whose establishment is forbidden, one of said bridges being considered to be the head-end bridge, said method comprising a bridge port attribute association step of the network during which is associated with the port, which provides access the most direct to said headend bridge, a root attribute, characterized in that it comprises

a step (200) for obtaining said associations between ports of each bridge and port attributes, and

a step (300) for defining the routing constraints relating to each bridge during which the establishment of any internal link between two ports of a bridge is forbidden if one of these two ports is not associated to the root attribute.

2) Method according to claim 1, said step of association of port attribute of a bridge being executed as soon as an event occurs on said local or infrastructure network, characterized in that said step (200) of obtaining said associations between ports of each bridge and port attributes is preceded by a step (100) for detecting the updating of the attributes associated with the ports of at least one of said bridges.

3) Method according to claim 1 or 2, characterized in that said step (200) for defining the routing constraints comprises a step of associating the root attribute associated with a port of a bridge with a so-called attribute of head (H) and each other attribute of a port of a bridge to a subscriber attribute (S), the establishment of an internal link between two ports of a bridge then being prohibited if one of these two ports is not associated with the header attribute.

4) Method according to one of claims 1 to 3, each bridge being associated with an identifier formed inter alia of a priority value, characterized in that the bridge which is considered to be the headend bridge is the bridge associated with the identifier whose priority value is the lowest among the priority values of the identifiers associated with the other bridges.

5) Method according to one of claims 1 to 4, said routing constraints being defined from a virtual network determined from said local area network or infrastructure, characterized in that a logical topology of said virtual network is then obtained following a convergence of an STP type protocol.

6) Method according to one of claims 1 to 5, characterized in that said bridges are Ethernet bridges.

A service flow routing constraint setting apparatus for associating with one of the bridges of a local area network or infrastructure subject to said routing constraints, said routing constraints defining those of the internal links of the network. this bridge, the establishment of which is forbidden, one of the said bridges of the network being determined as being the head-end bridge, the said bridge comprising means (MAP ₁ ) for associating an attribute with each port of a bridge of the said network and to the port of each bridge which provides the most direct access to said head bridge an attribute called root (R), characterized in that it comprises: - means (MOA) to obtain said associations between ports of each bridge and attributes of port, and

means (MDA) for defining routing constraints of said bridge by prohibiting the establishment of any internal link between two of its ports if one of these two ports is not associated with the root attribute (R) .

8) Ethernet bridge of a local area network or infrastructure consisting of bridges interconnected with each other and intended to be subject to service flow constraints, said routing constraints defining those of the internal links of this bridge whose establishment is prohibited, one of said bridges of the network being determined to be the head-end bridge, said bridge including means (MAP) for associating an attribute with each port of a bridge of said network and with the port of each bridge which provides the most direct access to said head bridge a so-called root attribute (R), characterized in that it is associated with a device according to claim 7. 9) Computer program for establishing service flow routing constraints of a bridge of a virtual local area network subject to said routing constraints, characterized in that it implements the steps of the method of establishment according to one of claims 1 to 6.