EP1766899A1

EP1766899A1 - Resource management device for a communication network with inter-node connections associated with shared resource preemption rights and access-to-resource priority

Info

Publication number: EP1766899A1
Application number: EP05776460A
Authority: EP
Inventors: Richard Douville; Martin Vigoureux
Original assignee: Alcatel CIT SA; Alcatel SA
Current assignee: Alcatel Lucent SAS
Priority date: 2004-07-08
Filing date: 2005-06-14
Publication date: 2007-03-28
Also published as: CN101006689A; FR2872981A1; US20080320153A1; WO2006013295A1; FR2872981B1

Abstract

The invention concerns a device (D) dedicated to the management of resources in a communication network comprising nodes (LER, LCR) mutually connected via connections. Said device (D) comprises management means (MG) for defining connections and associating same with resources such that said connections may be set up if needed. The management means (MG) are further configured to associate with each connection, based on a selected criterion, either a first status wherein it is provided with a preemption right on the selected resources shared at least partly with another connection, without being authorized to use them, or a second status wherein the connection is authorized to use the selected resources for setting up said connection.

Description

RESOURCE MANAGEMENT DEVICE FOR INTER-NODE CONNECTION COMMUNICATIONS NETWORK ASSOCIATED WITH SHARED RESOURCE PRE-FEED RIGHTS AND PRIORITY ACCESS TO RESSOURCES

The invention relates to the field of connection-based communications networks, and more particularly the devices responsible for managing the resources used by connections within such networks.

Here "connection" is understood to mean a circuit (physical or virtual) for data transmission, defined between two nodes of the network, with possible intermediate nodes interposed between these two nodes. A connection can therefore be seen as a single logical link between two nodes (whether they are neighbors or not).

In networks with connections supporting at least two switching layers, the so-called lower layer provides the so-called upper layer connectivity through connections established at its level. This connectivity allows the transport of data from the upper layer. For example, in an IP network following (G) MPLS type protocols (for

"(Generalized) MultiProtocol Label Switching"), information about the available connections is transmitted to the upper layer by means of a TE-type extension link-like routing protocol, such as the OSPF-TE protocol ( for Open Shortest Path First-Traffic Engineering). Thus, when the upper layer wants data to be transported, it first analyzes the information about the available connections to determine if one or a series of these available connections can carry the data, and if so, transmits them to the lower layer to manage their transport. As known to those skilled in the art, in the above-mentioned networks the connections are planned in advance in order to anticipate the needs of the upper layer or to prevent congestion phenomena occurring on the established connections. These connections can be statically established by estimating the present and future needs of network users. This has the disadvantage of permanently reserving network resources at the level of the data plan which will be exploited, for some, only episodically. This induces a temporal waste of resources. Another way to do this is to establish connections dynamically by estimating customer needs at each moment and reassigning connections based on a needs estimate. For example, the load of each established connection is analyzed, and when this load exceeds a chosen threshold, a new connection is established in order to anticipate a possible congestion. This is accomplished by reserving network resources at the data plane level for each provisioning connection, so that the connection can be established in the lower layer. This way of proceeding allows a temporal re-use of the resources. This induces a sizing of the number of resources required for the network less than that of a network operating with static connections.

The major drawback of this resource management mode lies in the fact that, when creating provisioning connections across the network, it is necessary to reserve a large number of resources for each of these connections, whereas these will not always be used by the upper layer in real operating conditions. The criterion for creating provisioning connections is related to an estimate of customer needs. This estimate can be improved, and is more so if we want to ensure a high level of quality of service to customers in order to ensure almost always new connections available for them (the extreme solution is to provide constantly connections and return to a static model). In other words, the network is oversized in number of resources and underutilized (effectively) and therefore more expensive than it should be. In an attempt to improve the situation, it has been proposed to use higher level load detection thresholds. However, in the presence of such thresholds oscillation phenomenon can occur in the network. In fact, the network is frequently obliged to release very quickly a Provisioning connection previously established, forcing the lower layer to signal the unavailability of this previous connection to the upper layer. This can lead to data transmission errors at the top layer, when the frequency of settlements and releases becomes too high. Indeed, all the nodes of the network are not informed instantly of the disappearance of a connection and can try to use it.

It was also proposed not to establish provisioning connections in advance. But in this case, on the one hand, the upper layer is not informed of the transport capacity of the lower layer and / or accessibilities through the lower layer, and secondly, the quality of service is weak because contentions can intervene at the level of customers and obtaining a new connection suffers from a more or less important delay. Since no known solution is entirely satisfactory, the purpose of the invention is therefore to improve the situation.

It proposes for this purpose a resource management device, for a communications network comprising nodes connected to each other by connections, and comprising management means responsible for defining the connections and associating them with resources so that these connections can be established when needed.

This device is characterized by the fact that its management means are also responsible for associating with each connection, according to a chosen criterion, a first state in which it has a right of pre-emption on selected resources shared at the same time. least partially with at least one other connection, without being authorized to use them, or a second state in which it is authorized to use the resources selected for the establishment of this connection.

The device according to the invention may comprise additional characteristics that can be taken separately or in combination, and in particular:

its management means can be responsible for ordering the transmission to the upper layer, by the lower layer, of the respective states of the connections, so that the upper layer knows the connections placed in their first state and can ask the lower layer that these connections are used for the transport of data, if necessary, its management means can be responsible for associating a preemption priority level associated with a connection,

• When resources are shared by connections associated with a second state and having different preemption priority levels, the connection with the highest priority level is preferentially prioritized for preemption of the shared resources when it wishes to establish this connection,

• in the presence of resources shared by first and second connections having different first and second preemption priority levels (the first level being lower than the second level), and when the first connection is in its second state while the second connection passes from its first state to its second state, the second connection is authorized to preempt the shared resources used by the first connection, in order to establish this other connection, the first connection established being then interrupted (or released),

Alternatively or in addition, in the presence of resources shared by first and second connections having different first and second preemption priority levels (the first level being always lower than the second level), and when the second connection is in its second state while the first connection goes from its first state to its second state, the first connection is not allowed to pre-empt the shared resources used by the second connection, in order to be established, as long as the second established connection remains in its second state. It is therefore interesting to set up or keep connections associated with a first state and recorded on at least one resource used by a connection associated with a second state having a preemption priority level higher than the his,

its management means may be responsible for associating with the connections associated with a first state a level of probability of establishment, for example depending on the load of connections which are at the same time established and used in the network, where resources are exclusively shared by connections associated with a first state having different levels of probability of establishment, the connection having the highest level of probability of establishment is preferentially priority for the configuration of the shared resources. If resources are reserved by a connection associated with a second state, then no other connection can change the configuration of those resources without having previously pre-empted that connection. This probability of establishment, the use of which is exclusively reserved when sharing resources between connections associated with a first state, makes it possible to reduce the delay in establishing a connection by passing from its first state to its second state. assuming that part of the resources that it uses has been pre-configured (gain on the configuration time) before its actual reservation, - its management means can be responsible for changing the state of a connection of the first state to the second state when the load of another established connection exceeds a chosen threshold, and / or when receiving a specific request from a user of the network, the exceeding of the threshold and the request constituting each a chosen criterion, its management means can be loaded, if it is necessary to change the state of a connection from the first state to the second state, to order the transmission to the nodes through which this onnion, a message of change of state of the connection intended to trigger a preemption of the resources with a view to its establishment, - its management means can be loaded, before associating a first state with a new connection, of order the preparation of a switching matrix between connections integrating this new connection, so that during the change of state of the new connection, the first state to the second state, the new connection can be established substantially immediately.

The invention is particularly well suited, although not exclusively, to circuit-switched networks with an IP control plane following (G) MPLS protocols.

Other features and advantages of the invention will appear on examining the detailed description below, and the attached drawing, in which the single figure schematically illustrates an example of a communications network with connections equipped with a resource management device according to the invention. The appended drawing may not only serve to complete the invention, but also contribute to its definition, if any.

The object of the invention is to make it possible to optimize the sizing of connection-based communications networks and to optimize the use of their resources. To this end, it proposes one or more resource management devices D intended to be implemented in a communications network with connections supporting at least two so-called upper and lower network protocol layers.

By "upper layer" is meant here a protocol layer responsible for processing data to be transported from one node of the network to another node of this network, and "lower layer" means a protocol layer responsible for managing transport over connections ( once established) data from the upper layer.

In what follows, it is considered as an illustrative example that the network is an IP (Internet Protocol) network following GMPLS (Generalized MultiProtocol Label Switching) protocols. But, the invention is not limited to this type of network. It also relates, for example, circuit switched networks (wavelengths, wavelength groups, or SONET / SDH) for the lower layer. In addition, it is considered in the following that the protocol used by the lower layer to transmit to the upper layer information relating to the connections is a TE-type extension link state routing protocol, such as the OSPF protocol. -TE (Open Shortest Path First-Traffic Engineering).

The IP GMPLS network illustrated in the single figure belongs to the family of so-called "tag-switched" communications networks (or

"Label Switched Network"). Such a network comprises network equipment (LERn, LCRm) defining nodes and each generally comprising a load distribution device (not shown).

Network devices are label switching routers (or LSRs) coupled to each other. Specifically, these LSRs can be grouped into two categories: Peripheral Routers (or LERs for "Label Edge

Routers ") LERn (here n = 1 to 4), and the core routers (or LCRs for

"Core Routers Label") LCRm (here m = 1 to 3). In particular, LER routers are responsible for establishing circuits (or switching paths (or

LSPs for "Label Switched Paths")), and LCR routers are responsible for switching data packets.

Here, the term "LSP (or virtual circuit)" is understood to mean a path between a source LER peripheral router and a destination peripheral LER router, defined by a sequence of links each established between two nodes. Furthermore, here is meant by "link" a physical connection between two routers LSR. In the single figure the links are materialized by two-way arrows.

It is recalled that an LSP path is usually calculated to optimize traffic transport between a source LER peripheral router and a destination LER peripheral router. In a GMPLS IP network, each LER peripheral router is arranged, when it is a source, to calculate the best LSP path for transferring the data streams it receives to the destination LER peripheral router, taking into account the service. associated with the feeds, the current topology of the network and the current loads of the links. It is important to note that the LER destination router of an LSP path is the last router to which a tagged packet is transmitted within a zone (or domain), but not the destination address of said tagged packet. Each LER peripheral router and each core router LCR has a switching matrix that is used to direct the data streams received in at least one established connection. Here, the term "established connection" means a connection ready to carry data streams due to the configuration and specific activation (configuration) of the switching matrices of the various LER and LCR routers that constitute it.

A multiplicity of user terminals or companies Ti (here i = 1 to 4) are likely to connect to at least some of the LERs routers to be able to exchange data between them.

The GMPLS IP network also includes a Network Management System (NMS) for transmitting data to the LSRs and extracting data from them to allow network management by an administrator (or manager). The NMS generally comprises an SG management server in which can be implemented a resource management device D according to the invention, which will be described below. In this case, the device D is of centralized type so as to manage at least part of the resources of the network. But, it is possible to envisage a variant in which resource management devices D can be distributed over the entire network. In this case, each LSR router has a management device D according to the invention, dedicated to its own resources and not to all the resources of the network (or a part of it), as in the case centralized.

In the centralized case, the device D comprises an MG management module responsible for defining the connections of at least part of the network and associating each of them, on the one hand, with selected resources, which can be shared at least partially with at least one other connection, and on the other hand, to a state representative of a right or a prohibition of pre-emption of the resources associated with them. More specifically, each connection may, depending on the circumstances that depend on at least one chosen criterion, be associated with a first state in which it has a right of preemption over the resources associated with it and that it shares with least partially with at least another connection, but without being authorized to use them, is associated with a second state in which it has the same right of preemption but is authorized to use the resources chosen for its establishment.

Thus, according to the invention, provision is made for the connections in advance, but, instead of associating specific resources with each of them, resources are shared between the connections, and each one is authorized or forbidden. connection to use the resources associated with it, and that it shares, in order to be established or relaxed (or not established), according to the needs of the moment. Preemption consists of reserving the resources associated with a connection at what is generally called the data plane. It can be noted that in an optical network the reservation of resources results in the activation of a cross connection (or "cross-connection") at the optical switch OXC. The criterion that is used by the management module MG to decide on the state that must be associated with each connection may for example be a comparison to an established connection load threshold. In this case, the management module MG is arranged to analyze the load of each established connection in order to compare it to the chosen threshold. Then, if the analyzed load (of one or more connections established) exceeds the threshold chosen, the management module MG orders the activation of switching matrices nodes belonging to a second connection not yet established so that the next flows of data are switched from the first connection to the second connection. This avoids congestion of the first connection.

Of course, other criteria can be used, either in place of the one presented above (threshold comparison), or in parallel. Thus, a criterion may consist in receiving or not receiving a specific request from a user (client) of the network. In this case, when the management module MG receives a specific request, it determines the connection associated with a first state and best suited to the needs of the customer, then it changes its state from the first state to the second state.

The activation of switching matrices is accompanied by the changing the state of the second connection from the first state to the second state.

Moreover, when at least one of the switching matrices to be activated has not been pre-configured to enable the switching of the data flows to the second connection (and therefore its local establishment), provision is made before the phase of activation a configuration phase that is performed when reserving resources for this connection.

When the state of the second connection has been modified and therefore it can preempt the resources associated therewith, the lower protocol layer transmits to the upper protocol layer a message intended to indicate to it that it has at its disposal a new connection established. The upper protocol layer can then transmit to the lower protocol layer, if it deems it necessary, the data to be transported by the new established connection. Preferably, the management module MG is also responsible for associating a priority level with the preemption right which is associated with one (and preferably each) connection. This makes it possible to prioritize the use of resources for certain connections.

For example, when resources are shared by connections with different priority levels, the network can be configured so that the connection with the highest priority is prioritized for preemption of the resources it shares with one or more other connections each time it wishes to be established. In other words, if two connections sharing resources must be established substantially simultaneously, only the one with the highest priority level will be able to use the resources associated with it and therefore be established. Similarly, if a first connection wishing to establish itself is associated with a preemption priority level higher than that of a second connection already established, then this first connection preempts the resources used by this second connection and the second connection is interrupted (or released).

It is important to remember that a first connection can only preempt the resources of a second established connection first connection is in a second state or goes from a first state to a second state.

On the other hand, when resources are shared by a first connection, having a first priority level with the preemption right and associated with a second state, and a second connection having a second priority level, higher than the first level, and associated to a state passing from the first state to the second state, then the network may be configured so that the second connection is allowed to preempt the shared resources that are used by the first established connection, so that it can be established in place of the first connection.

In other words, if an established connection, (and therefore associated with a second state) and associated with a weak preemption priority level, another connection, with which it shares resources, may cause its release, in order to be established in its place, since its priority level is higher than its own.

In addition, when resources are shared by a first connection, having a first priority level to the right of preemption and associated with a state passing from the first state to the second state, and by a second connection having a second priority level, higher than the first level, and associated with a second state, then the network may be configured so that the first connection is not allowed to preempt the shared resources that are used by the second established connection, so that it can be established instead the second connection, as long as the latter remains in its second state. Therefore, when it is desired to provision connections, it is inappropriate to set up a first connection associated with a first state and registered on at least one resource reserved by a second connection associated with a second state having a second state. preemption priority level higher than his own. On the other hand, this operation can be interesting in the framework of a reconfiguration of the connections of the network.

In other words, if an established connection (and thus associated with a second state) is associated with a high preemption priority level, another connection, with which it shares resources and newly associated with a second state, can not cause its release, to be established in its place, provided that the established connection goes into the first state.

It is important to note that preemption priority levels are not necessarily associated with connections permanently. It can indeed be envisaged that the management module MG is for example arranged so as to modify the priority level associated with one or more connections according to the needs of the moment or for a chosen duration. The management module MG may also be responsible for associating with at least some of the connections a level of probability of establishment which may for example be a function of the connection load established in the network.

For example, it is assumed that only one connection is established across the network to carry data between terminals T 1 and T 4. If this single connection arrives at saturation (and therefore has a high load), the network is more likely to establish a new connection between terminals T1 and T4 than if this connection was unfilled (and therefore low). The same probabilistic reasoning can be done in the presence of several connections of the network. Thus, between two connections, for example defined between the terminals T1 and T4 and between the terminals T1 and T3, one of the two connections can be more loaded than the other. If it is desired to set up a connection associated with a first state for each of these connections as a provisioning connection, then each of these provisioning connections can be associated with a probability of establishment reflecting the load level of the connections. they will have to make up for it. Thus, if these two provisioning connections share at least one common resource not reserved by a third connection associated with a second state, the connection having the greatest probability of establishment to the ability to configure the resource or resources concerned.

Finally, the management module MG can be loaded, when it decides to change the state of a connection from the first state to the second state, to instruct the lower protocol layer to transmit to each node, through which this connection passes, a status change message intended to trigger the preemption of resources so that it can be established. This is particularly advantageous because in a conventional network the reservation of resources at the nodes of a connection requires a first pass from the source node to the destination node to ask each node if it has available resources, then a second pass of the node destination to the source node to effectively reserve from each node the resources requested during the first pass. The management device D according to the invention, and in particular its management module MG, can be implemented in the form of electronic circuits, software (or computer) modules, or a combination of circuits and software.

The invention is not limited to the resource management device embodiments described above, by way of example only, but encompasses all variants that the person skilled in the art may envisage in the context of the claims. below.

Thus, in the foregoing, an exemplary implementation of the invention in which the management device was centralized so as to manage all or part of the resources of the network is described. But, resource management can also be distributed across the entire network. In this case, each node (or router) has a management device according to the invention, dedicated to its own resources.

Claims

A resource management device (D) for a communications network comprising nodes (LER, LCR) interconnected by connections, said device comprising management means (MG) suitable for defining said connections and associating them with resources such that said connections can be established when necessary, characterized in that said management means (MG) are further arranged to associate each connection, according to a selected criterion, a state from a first state wherein said connection has a preemptive right on selected resources shared at least partially with at least one other connection, without being allowed to use them, and a second state in which said connection is allowed to use said selected resources for of the establishment of said connection.

2. Device according to claim 1, characterized in that said management means (MG) are arranged, before associating a first state with a new connection, to order the preparation of a switching matrix between connections incorporating said new connection so that in case of a change of state of said new connection, from the first state to the second state, said new connection can be established substantially immediately.

3. Device according to one of claims 1 and 2, characterized in that said management means (MG) are arranged to order the transmission to a so-called upper network protocol layer by a so-called lower network protocol layer, managing the transport by said connections, once these connections are established, data from said upper layer, respective states of the connections, so that said upper layer knows the connections placed in their first state and can ask said lower layer that these connections are used to data transport, if needed.

4. Device according to one of claims 1 to 3, characterized in that said management means (MG) are arranged to associate a priority level preemption right associated with a connection.

5. Device according to claim 4, characterized in that in the case of resources shared by connections having priority levels associated with a second state and having different preemption priority levels, the connection having the priority level. the highest priority for preemption of said shared resources, to establish said connection.

6. Device according to claim 5, characterized in that in the presence of resources shared by first and second connections having different first and second preemption priority levels, said first level being lower than said second level, and when said first connection is associated with a second state while the second connection moves from the first state to the second state, said second connection is allowed to pre-empt said shared resources used by said first connection to establish said connection, said first connection being then interrupted.

7. Device according to one of claims 5 and 6, characterized in that in the presence of shared resources by first and second connections having different first and second priority levels preemption right, said first level being lower than said second level, and when said second connection is associated with a second state while the first connection goes from the first state to the second state, said first connection is not allowed to preempt said shared resources used by said second connection, in order to establish said connection, as long as said second established connection remains associated with a second state.

8. Device according to one of claims 1 to 7, characterized in that said management means (MG) are arranged to associate with at least said connections, associated with a first state, a level of probability of establishment.

9. Device according to claim 8, characterized in that said level of probability of establishment is a function of the connection load established and used in said network.

10. Device according to one of claims 1 to 9, characterized in that said management means (MG) are arranged to change the state associated with a connection from the first state to the second state when the load of another established connection exceeds a chosen threshold, the threshold exceeding constituting said selected criterion.

1 1. Device according to one of claims 1 to 10, characterized in that said management means (MG) are arranged to pass the state associated with a connection of the first state to the second state upon receipt of a request of a user, said request constituting said selected criterion.

12. Device according to one of claims 1 to 1 1, characterized in that said management means (MG) are arranged, in case of need of change of state of a connection of the first state to the second state, for order the transmission to the nodes (LER, LCR), through which said connection passes, a change of state message of said connection intended to trigger a preemption of resources for the establishment of said connection.

13. Router (LER, LCR) for a network communications connections, characterized in that it comprises a management device (D) according to one of the preceding claims, dedicated to the management of its own resources.

Management server (SG) for a communications network comprising nodes (LER, LCR) interconnected by connections, characterized in that it comprises a management device (D) according to one of claims 1 to 12, dedicated to managing at least part of the resources of said network.

15. Use of the resource management device (D) according to one of claims 1 to 12 in IP networks following protocols (G) MPLS.

16. Use of the resource management device (D) according to one of claims 1 to 12 in circuit-switched networks.