Assistance with network planning

Abstract

Assistance in capacity planning of a network The invention relates to a method for assisting in capacity planning in a network comprising points of presence connected by backbone links, including for each point of presence: - determining (205) significant backbone links as a function of the respective contributions and a minimum contribution criterion; - identifying (206) a weakest backbone link as the significant backbone link for which a maximum additional bandwidth value that can be added to the point of presence without saturating the significant backbone link is the lowest. Figure for abstract: [Fig 2]

Description

Title of the invention: Assistance in planning a network Technical field

[0001] The invention relates to the field of planning a data transport network.

[0002] It relates in particular to the planning of an IP network, notably in a backbone network. State of the art

[0003] An IP backbone network is a data transmission network that carries a large portion of global IP traffic and uses fast transmission technologies and high bandwidths. It is therefore the nerve center of the IP network.

[0004] A backbone network makes it possible to interconnect IP networks specific to remote geographical areas. The backbone network interconnects, in particular, fixed operator networks, such as Wi-Fi, FTTX, local networks and mobile networks, such as 3G, 4G, 5G or any other generation.

[0005] For this purpose, according to a PoP topology, for "Point of Presence" in English, points of presence are interconnected in the network via backbone links. Each point of presence can be seen as one or more routers configured to collect, transport and route data flows, from point of presence to point of presence, towards their final destination. The routers are often located in the same room or data center, but can be distributed in several locations. In addition, some points of presence are access points allowing networks or communicating devices to access the backbone network, and thus to communicate with other networks and communicating devices located at a great distance.

[0006] The planning of such a network may require the definition, and adaptation, of routing metrics, such as for example ISIS metrics of an IP network, in order to guarantee the best use of a given infrastructure for the transport of data flows via the backbone links of the network, while ensuring resilience of the network to cases of failures on one or more backbone links.

[0007] Dynamic adaptation of routing metrics is, however, complex to implement in practice.

[0008] For this reason, an operator in charge of network planning and dimensioning generally works with relatively constant routing metrics (i.e. with regular, but infrequent updates), and must therefore regularly monitor the network to detect overloaded backbone links or network points of presence that require additional capacity (to anticipate new data traffic demand and / or to ensure redundancy in the event of a network failure), or conversely to request a reduction in the size of the backbone links that connect it to the network in the event that the traffic of the point of presence has decreased.

[0009] For network monitoring, the operator generally has at his disposal the load status of the backbone links in a nominal situation as well as the respective capacities of the backbone links. The operator thus bases network monitoring on the load status of the backbone links between the points of presence, in the nominal situation. However, this data alone does not allow the full complexity of the network to be understood, particularly in the event of a breakdown, and offers an incomplete view of the network.

[0010] There is thus a need to improve assistance in planning the capacities of a network, by identifying the backbone link(s) of the network on which the operator's attention must be focused, while taking into account the complexity of the network.

[0011] The invention offers a solution which does not have the drawbacks of the state of the art. Statement of the invention

[0012] To this end, according to a functional aspect, the invention relates to a method for assisting in capacity planning in a network comprising a plurality of points of presence connected to each other in the network by links, called backbone links, the method comprising the following steps, implemented for at least one of the points of presence:

[0013] - determination of significant backbone links, among the backbone links connecting said at least one point of presence in the network, based on the respective contributions of the backbone links to the transport of data from and / or to said at least one point of presence; - identification of a weakest backbone link among the significant backbone links, the weakest backbone link being the significant backbone link for which a maximum additional bandwidth value that can be added to said at least one point of presence without saturating the significant backbone link, is minimum compared to the maximum additional bandwidth values ​​that can be added to said at least one point of presence without saturating the other significant backbone links respectively.

[0014] Thus, the method according to the invention makes it possible to identify the most critical backbone link(s) for a point of presence, which is a limiting backbone link for the additional bandwidth that can be allocated to the point of presence. The operator is thus helped in network planning.

[0015] Furthermore, the maximum additional bandwidth value associated with the weakest backbone link is useful for the operator in that it corresponds to a level of bandwidth available for sale for the point of presence. The weakest backbone link is thus the link whose capacity is to be increased if the bandwidth available for sale is considered too low, taking into account future connection requests likely to appear for this point of presence.

[0016] According to embodiments, the maximum additional bandwidth values ​​may be determined from respective capacities and load values ​​of the significant backbone links.

[0017] Thus, the maximum additional bandwidth values ​​can be determined from dynamic data such as the respective load values ​​of the significant backbone links, thereby improving network planning.

[0018] According to embodiments, the method may further comprise, based on a comparison between the maximum additional bandwidth value that can be added to said at least one point of presence without saturating the weakest backbone link, and a minimum available bandwidth criterion, generating an alert identifying the weakest backbone link for said at least one point of presence.

[0019] Thus, this embodiment makes it possible to alert the operator only in the case where the additional bandwidth value is too low. It is thus possible to limit the number of alerts generated for the operator, while reinforcing the relevance of these alerts. Indeed, in the prior art, it may be provided to compare the load in a nominal situation of each backbone link with a ratio of the capacity of the link, and to generate an alert each time the load in a nominal situation exceeds the ratio, which leads to generating a very large number of alarms, the few relevant alarms then being able to go unnoticed.

[0020] According to embodiments, the method may further comprise obtaining respective load values ​​of the backbone links connecting said at least one point of presence to the network, the contributions of the backbone links to the transport of data from and / or to said at least one point of presence being determined from the respective load values ​​of the backbone links connecting said at least one point of presence to the rest of the network.

[0021] Thus, the contributions can be determined from measured and / or simulated data and evolving dynamically, which makes it possible to improve the determination of significant backbone links.

[0022] In addition, obtaining respective load values ​​of the backbone links connecting said at least one point of presence to the network may comprise: - obtaining an average load value over a given period for any backbone link among the backbone links connecting said at least one point of presence to the network, the average load values ​​being associated with a nominal situation; and / or - obtaining a peak load value over a given period for any backbone link among the backbone links connecting said at least one point of presence to the network, the peak load values ​​being associated with a peak load situation; and / or - obtaining, from a network simulator, for any given backbone link among the backbone links connecting said at least one point of presence to the network, a load value associated with a worst-case failure situation in the network for the given backbone link.

[0023] Thus, the respective load values ​​obtained correspond to the same situation, or to the same set of situations, and are therefore comparable, which improves the precision of the method according to the invention, in particular the determination of the significant backbone links and therefore the determination of the weakest backbone link.

[0024] In addition, for any given backbone link connecting said at least one point of presence to the network, at least one contribution of this link to the transport of data from and / or to said at least one point of presence can be determined in at least one situation among the nominal situation, the peak load situation and the worst case failure situation for the given backbone link.

[0025] Thus, the accuracy associated with determining significant backbone links is improved.

[0026] In addition, for any given backbone link connecting said at least one point of presence to the network, a first contribution to the transport of data from and / or to said at least one point of presence can be determined for the nominal situation, a second contribution to the transport of data from and / or to said at least one point of presence can be determined for the peak load situation, and a third contribution to the transport of data from and / or to said at least one point of presence can be determined for the worst case failure situation for the given backbone link, and the given backbone link can be identified as significant based on the three determined contributions.

[0027] Thus, the accuracy associated with the determination of significant backbone links is improved. It is furthermore made possible to take into account several given situations, whereas the solutions of the prior art only consider the nominal situation. It is thus possible to account for the complexity of the network, in particular on rerouting situations in the event of a failure.

[0028] In addition, for any given backbone link connecting said at least one point of presence to the network, the first contribution can be compared to a first minimum contribution threshold, the second contribution can be compared to a second minimum contribution threshold, and the third contribution can be compared to a third minimum contribution threshold, and the given backbone link can be identified as significant based on results of the three comparisons.

[0029] Thus, differentiated criteria for each of the situations can be applied for the determination of significant backbone links.

[0030] Additionally, the given backbone link can be determined as significant only if the first contribution is greater than the first minimum contribution threshold, the second contribution is greater than the second minimum contribution threshold, and the third contribution is greater than the third minimum contribution threshold.

[0031] Thus, the significant backbone links ensure a minimum proportion of data traffic transport to and / or from the point of presence, in several given situations.

[0032] In addition or as a variant, the weakest backbone link of said at least one point of presence is the significant backbone link for which a maximum additional bandwidth value that can be added to said at least one point of presence without saturating the significant backbone link in the worst case failure situation for the significant backbone link, is minimal compared to the maximum additional bandwidth values ​​that can be added to said at least one point of presence without saturating respectively the other significant backbone links in the respective worst case failure situations of the other significant backbone links.

[0033] Thus, the maximum additional bandwidth value that is available for sale for the point of presence is determined in the worst case of failure, which makes it possible to ensure better resilience of the network in cases of failure on backbone links.

[0034] In addition, the maximum additional bandwidth value that can be added on uledit at least one point of presence without saturating a given significant backbone link in the worst case failure situation of the given significant backbone link can be determined from: - the capacity of the given significant backbone link; - the load value in the worst case of failure of the given significant backbone link; and - the contribution of the given significant backbone link for said at least one point of presence in the worst case of failure of the given significant backbone link.

[0035] Thus, the accuracy associated with determining the weakest backbone link and the associated additional available bandwidth value is improved.

[0036] According to embodiments, the steps of the method can be applied to at least one other point of presence of the network.

[0037] Thus, an operator can adapt the capacity planning in the network, or in a part of the network, considering the points of presence of the network, or of the part of the network, one after the other. This allows for an overall optimization of the capacities of the network or of the part of the network.

[0038] In addition, a backbone link identified as being insignificant for said at least one point of presence as well as for another point of presence connected to said at least one point of presence via said backbone link is identified as being a 'non-useful' backbone link for the network.

[0039] Thus, the method makes it possible to detect backbone links which are 'not useful' in transporting traffic in the network, which makes it possible to optimize operational costs and thus improve network capacity planning.

[0040] According to a hardware aspect, the invention also relates to a device for assisting in capacity planning in a network comprising a plurality of points of presence connected to each other in the network by links, called backbone links, the device comprising a processor configured, for at least one point of presence, to:

[0041] - determining significant backbone links, among the backbone links connecting said at least one point of presence in the network, depending on the respective contributions of the backbone links to the transport of data from and / or to said at least one point of presence; - identify, among the significant backbone links, a weakest backbone link, the weakest backbone link being the significant backbone link for which a maximum additional bandwidth value that can be added to said at least one point of presence without saturating the significant backbone link, is minimum compared to maximum additional bandwidth values ​​that can be added to said at least one point of presence without saturating the other significant backbone links respectively.

[0042] According to another material aspect, the invention also relates to a computer program capable of being implemented on a device such as aforementioned, the program comprising code instructions which, when the program is executed by a processor, carry out the steps of the defined method.

[0043] Such programs may use any programming language. They may be downloaded from a communications network and / or recorded on a computer-readable medium.

[0044] According to another material aspect, the invention relates to a data medium on which at least one series of program code instructions has been stored for the execution of the method defined above. Brief description of the drawings

[0045] The invention will be better understood on reading the following description, given by way of example and with reference to the appended drawings in which:

[0046] [Fig. 1] illustrates a network connecting points of presence by so-called backbone links, according to embodiments of the invention.

[0047] [Fig.2] is a method of generating an alert to assist in network capacity planning, according to embodiments of the invention.

[0048] [Fig.3] illustrates a graph representing the additional bandwidth values ​​available to a point of presence and permitted by the backbone links linking the point of presence to the network;

[0049] [Fig.4] illustrates the structure of a network capacity planning assistance device, according to embodiments of the invention. Description of the embodiments

[0050] [Fig.l] illustrates a network 100 interconnecting points of presence 101.1 to 101.8. The network 100 may be a backbone network, of the IP type for example, capable of interconnecting points of presence respectively located in places distant from each other, for example in different countries or continents.

[0051] The network 100 interconnects a first point of presence 101.1, a second point of presence 101.2, a third point of presence 101.3, a fourth point of presence 101.4, a fifth point of presence 101.5, a sixth point of presence 101.6, a seventh point of presence 101.7 and an eighth point of presence 101.8. The points of presence are identified by a letter from A to H, as shown in [Fig.l].

[0052] Eight points of presence 101.1 to 101.8 are shown in [Fig.l]. However, the network 100 may comprise several tens or even several hundreds of points of presence, no restriction being attached to the number of points of presence of the network 100.

[0053] In order to interconnect the points of presence 101.1 to 101.8, the network 100 comprises backbone links, each backbone link being defined by: - a capacity or size of the backbone link, expressed as a quantity of data per unit of time, for example in bytes per second, and which corresponds to the maximum bandwidth that the link can achieve for the transfer of data between the points of presence that it connects. Note that in an IP backbone network, the capacity of a link can be of the order of several megabytes per second, or even several tens or hundreds of megabytes per second, or even several gigabytes per second. For the sake of simplification in what follows, bandwidths can be expressed as quantities of data, without reference to a unit of time; - a nominal or average load value, which is the average value of the band bandwidth used for data transport on the backbone link, in other words the average value of the backbone link load. For example, the nominal or average load can be an average value of the bandwidth used on the link over a given period, of several days and preferably several weeks; - a peak load value, which can be determined from at least one local maximum of the bandwidth used for transporting data on the backbone link, in other words the local maximum of the load of the backbone link. The peak load value can be an average value of the local maxima, or of some of the local maxima, of the bandwidth on the backbone link over a given period, for example a period of several days and preferably several weeks; - routing metrics, used for routing data in the backbone network, in particular for determining a route comprising intermediate points of presence between a source point of presence and a destination point of presence of data. The routing protocol used in the network 100 may be an IS-IS protocol, which is a link-state protocol. Alternatively, the OSPF protocol, for “Open Shortest Path First”, which is also a link-state protocol, may be used in the network 100.

[0054] The peak and average load values ​​can be accessed for each link of the network 100 by measuring the quantity of data, at a given measurement step, for example by means of probes in the network 100, and by storing the evolution of the bandwidth or load.

[0055] Each point of presence may be a single server, or a network infrastructure comprising a plurality of network entities. In the example of [Fig.l], the first point of presence 101.1 is a network infrastructure, organized according to a given topology, the network entities being connected to each other by internal links to the point of presence. An internal link is thus a link connecting two network entities of the same point of presence, while a backbone link connects two distinct points of presence to each other via the backbone network 100. As shown in [Fig.l], the first point of presence 101.1 comprises a plurality of internal links. No restriction is attached to the topology of a point of presence, according to the invention.

[0056] In the context of the invention, a point of presence may comprise any number of backbone links. The network 100 may in particular comprise several hundred backbone links connecting two distinct points of presence.

[0057] The first point of presence 101.1 is connected to the network 100 by several backbone links, namely: - a first backbone link 110.1 with the second point of presence 101.2; - a second backbone link 110.2 with the third point of presence 101.3; - a third backbone link 110.3 with the fourth point of presence 101.4; - a fourth backbone link 110.4 with the fifth point of presence 101.5; and - a fifth backbone link 110.5 with the sixth point of presence 101.6.

[0058] The network 100 comprises other backbone links 111.1 to 111.7 which connect the points of presence 101.2-101.8 other than the first point of presence 101.1, namely: - another backbone link 111.1 between the second point of presence 101.2 and the third point of presence 101.3; - another backbone link 111.2 between the third point of presence 101.3 and the eighth point of presence 101.8; - another backbone link 111.3 between the eighth point of presence 101.8 and the seventh point of presence 101.7; - another backbone link 111.4 between the fourth point of presence 101.4 and the eighth point of presence 101.8; - another backbone link 111.5 between the sixth point of presence 101.6 and the seventh point of presence 101.7; - another backbone link 111.6 between the fifth point of presence 101.5 and the seventh point of presence 101.7; and - another backbone link 111.7 between the fifth point of presence 101.5 and the fourth point of presence 101.4.

[0059] [Fig.2] is a diagram illustrating the steps of a method of generating an alert to assist in the planning of a network, such as the network 100 illustrated in [Fig.l], according to embodiments of the invention.

[0060] The described method can be implemented in a network planning assistance device, such as the device 400 described in the following with reference to [Fig.4]. No restriction is attached to the location of the planning assistance device, which can be hosted in a network platform, such as a cloud-type platform for example.

[0061] In a step 200, the method is initialized, for example by starting the device 400 by an operator in charge of planning at least part of the network 100, in particular in charge of a subset of points of presence of the network 100, comprising at least one point of presence (for example the first point of presence 101.1 described previously), and able to comprise all of the points of presence of the network 100

[0062] At a step 201, the device 400 determines whether there is at least one point of presence in the subset for which the device 400 is responsible, to which to apply the following steps 203 to 210. If this is not the case, i.e. if steps 203 to 210 have been applied to the subset of points of presence, then the method ends at a step 202.

[0063] If this is not the case, i.e. if the device 400 determines that there is at least one remaining point of presence of the subset to which steps 203 to 210 have not been applied, the device 400 selects one of said at least one remaining point of presence for application of the following steps to the selected point of presence. It is considered in the following that the device 400 is in charge of the first point of presence 101.1 and the third point of presence 101.3, for illustrative purposes only. During a first iteration of the method, the device 400 may select the first point of presence 101.1.

[0064] At a step 203, the device 400 obtains, for any backbone link 110.1 to 110.5 connecting the first point of presence 101.1 to the network 100, the following information: - the capacity of the backbone link; - the average load value on the backbone link and / or the peak load value on the backbone link.

[0065] Such information may be stored in a memory of the device 400 and / or obtained from an entity external to the device 400. For example, the capacities of the links being static data, unless changes are made to the data transport infrastructure in the network 100, may be stored in a memory of the device 400 or in a memory directly accessible by the device 400. The average and peak load values ​​may be obtained from an external entity, in charge of collecting traffic data on the network 100, or from / to the points of presence of the subset only. The external entity may collect point load values ​​and may determine the average value and the peak value by accumulating point values ​​over a given period, such as the aforementioned period of several days or several weeks.In this case, the external entity communicates to the device 400 such average and peak values, for all the backbone links connecting the first point of presence 101.1 to the network 100. Alternatively, the device 400 accesses the point load values ​​of the links and determines the peak and average values ​​from the accumulation of point values ​​over a given period, such as the aforementioned period of several days or several weeks.

[0066] The device 400 can also obtain, in step 203 or in a prior step, descriptive data of the topology of the network, in particular descriptive of all the routers forming the points of presence and of the backbone links between these routers. In other words, the device 400 obtains the description of the network, consisting of all the points of presence and containing for each point of presence, the list of routers and the description of the backbone links which connect the routers together. The respective locations of the routers are also obtained, which makes it possible to determine the routers which belong to the same point of presence.

[0067] At a step 204, following step 203, the device 400 obtains, for example by means of a network simulator, for all the backbone links 110.1 to 110.5 connecting the first point of presence 101.1 to the network 100, the following information, for a given backbone link among the backbone links 110.1 to 110.5: - a load value of the given backbone link in the worst case failure. The worst case failure can be obtained from a network simulator, capable of simulating a set of failure cases in the network, and determining the failure case inducing the highest load value for the given backbone link, identified as the worst case failure for the given backbone link. The adjective "worst" is relative to the given backbone link, and does not indicate that the failure case is the worst in general for the network. The set of failure cases simulated in the network can correspond to 'single' failure cases (where a single backbone link in the network is faulty, the single faulty backbone link being any backbone link other than the given backbone link for which the worst case failure is determined), or to 'multiple' failure cases in which the simultaneous failures of several backbone links are considered.In the single failure case, if the network comprises N backbone links, Nl failure cases are simulated for the Nl backbone links other than the given backbone link for which the worst failure case is determined. Alternatively, the multiple failure cases concern a set of failure cases on several backbone links other than the given backbone link. Each simulated failure case induces an increase in load on the given backbone link, and ultimately allows obtaining for each backbone link: . - the maximum load value induced on this link, which is the maximum load corresponding to the worst case of failure; - the identification of the worst-case failure scenario, i.e. an identifier of at least one failed backbone link (other than the given backbone link) for single failure cases; or the identifiers of several backbone links assumed to be in simultaneous failure for multiple failure cases. It is thus understood that the worst-case failure is a relative expression, in that the worst-case failure for a first backbone link among the backbone links 110.1 to 110.5, may be different from the worst-case failure for a second backbone link among the backbone links 110.1 to 110.5.

[0068] Thus, following steps 203 and 204, the device 400 has the load values ​​of the backbone links 110.1 to 110.5, for each of the following situations: - nominal situation corresponding to the average load value of a given backbone link; - peak situation corresponding to the peak load value of the given backbone link; - worst case failure situation corresponding to the load value of the given backbone link for the worst case failure determined for this given backbone link. It is recalled that the worst case failures can differ between the different backbone links 110.1 to 110.5. For example, the worst case failure for the first backbone link 110.1 may correspond to a failure of the other link 111.6 while the worst case failure for the second backbone link 110.2 may correspond to a failure of the other link 111.4.

[0069] Note that, according to embodiments, the incoming traffic and the outgoing traffic of the first point of presence may be considered separately. In this case, the device 400 obtains in step 203 an average incoming load value and an incoming peak load value for any backbone link among the backbone links 101.1 to 101.5, and an average outgoing load value and an outgoing peak load value for any backbone link among the backbone links 101.1 to 101.5. Similarly, the device 400 can obtain, during step 204, for any given backbone link among the backbone links 101.1 to 101.5, a worst-case failure incoming load value for the backbone link and a worst-case failure outgoing load value for the given backbone link.

[0070] In a step 205, the device 400 determines, for at least one of the three situations described previously, and preferably for the three situations described previously, a contribution of each backbone link 110.1 to 110.5 to the transport of incoming traffic to and / or outgoing from the first point of presence 101.1, the contribution being a rate of contribution to the transport of traffic to and / or from the first point of presence 101.1. Alternatively, the contribution is not expressed as a rate or a percentage but as an absolute value of the traffic load transported from and / or to the first point of presence 101.1.

[0071] When differentiated load values ​​for the incoming and outgoing traffic are obtained in steps 203 and 204, each contribution of a given backbone link, determined in step 205, may be, in a given situation, a contribution of the given backbone link to the incoming and outgoing traffic of the first point of presence 101.1.

[0072] In a given situation, a total data load is exchanged by the first point of presence 101.1 with the other data points 101.2, 101.3, 101.4, 101.5 and 101.6. The total load is distributed among the different links, and the contribution of a given link among the backbone links 110.1 to 110.5, in a given situation, may be the fraction of the total load of the first point of presence 101.1 that is carried on the given backbone link.

[0073] Thus, the device 400 can determine, in the nominal situation, a first total load of the first point of presence 101.1 and: - a contribution from the first backbone link 110.1, corresponding to the share of the first total load transported on the first backbone link 110.1; - a contribution from the second backbone link 110.2, corresponding to the share of the first total load transported on the second backbone link 110.2; - a contribution from the third backbone link 110.3, corresponding to the share of the first total load transported on the third backbone link 110.3; - a contribution from the fourth backbone link 110.4, corresponding to the share of the first total load transported on the fourth backbone link 110.4; - a contribution from the second backbone link 110.1 corresponding to the share of the first total load transported on the fifth backbone line 110.5.

[0074] The device 400 can also determine, in the peak situation, a second total load of the first point of presence 101.1 and:

[0075] - a contribution from the first backbone link 110.1, corresponding to the share of the second total load transported on the first backbone link 110.1; - a contribution from the second backbone link 110.2, corresponding to the share of the second total load transported on the second backbone link 110.2; - a contribution from the third backbone link 110.3, corresponding to the share of the second total load transported on the third backbone link 110.3; - a contribution from the fourth backbone link 110.4, corresponding to the share of the second total load transported on the fourth backbone link 110.4; - a contribution from the second backbone link 110.1 corresponding to the share of the second total load transported on the fifth backbone link 110.5.

[0076] In addition or as a variant, the device 400 can also determine, or obtain from the external entity described above, in the respective worst-case failure situations of the different backbone links 110.1 to 110.5: - a contribution from the first backbone link 110.1, corresponding to the share of the total load of the first point of presence 101.1, transported on the first backbone link 110.1, in the worst case failure identified for the first backbone link; - a contribution from the second backbone link 110.2, corresponding to the share of the total load of the first point of presence 101.1 transported on the second backbone link 110.2, in the worst case failure identified for the second backbone link 110.2; - a contribution from the third backbone link 110.3, corresponding to the share of the total load of the first point of presence 101.1 transported on the third backbone link 110.3, in the worst case failure identified for the third backbone link 110.3; - a contribution from the fourth backbone link 110.4, corresponding to the share of the total load of the first point of presence 101.1 transported on the fourth backbone link 110.4, in the worst case failure identified for the fourth backbone link 110.4; - a contribution from the fifth backbone link 110.5 corresponding to the share of the total load of the first point of presence 101.1 transported on the fifth backbone link 110.5, in the worst case failure identified for the fifth backbone link 110.5.

[0077] The contributions of the backbone links 110.1 to 110.5, in the worst-case failure situations corresponding to them respectively, are obtained by the device 400 from the data resulting from the failure case simulations, obtained during the step 204 described previously. Alternatively, the device 400 obtains the contributions directly from the network simulator from which the load values ​​of each backbone link in its worst-case failure situation are derived.

[0078] In step 205, the device 400 determines the significant backbone links for the first point of presence 101.1, from among the set of backbone links 110.1 to 110.5, as a function of their contributions in at least one of the situations identified above, and as a function of a minimum contribution criterion. In addition, the device 400 determines the significant backbone links for the first point of presence 101.1, from among the set of backbone links 110.1 to 110.5, as a function of triplets of contributions obtained for the three situations described above.

[0079] Several embodiments can be provided for step 205, according to the invention.

[0080] In a first embodiment, the contributions of the backbone links 110.1 to 110.5 in a single given situation are taken into account, and the minimum contribution criterion is that a backbone link is significant if its contribution in the given situation is greater than a predetermined threshold. The given situation may be the nominal situation for example. Thus, the contribution in the given situation (nominal for example) of a given backbone link among the backbone links 110.1-110.5 is compared to the predetermined threshold, and if the contribution is greater than or equal to the predetermined threshold, the given backbone link is determined to be significant. Otherwise, the given backbone link is considered not to be significant or to be non-significant.

[0081] In a second embodiment, the contributions of the backbone links 110.1 to 110.5 in several given situations are taken into account, for example in the three aforementioned situations, and the minimum contribution criterion is that a backbone link is significant if a weighted sum or an average of its contributions in the given situations are greater than a predetermined threshold. Thus, the average or weighted sum of the contributions of a given backbone link among the backbone links 110.1-110.5 is compared to the predetermined threshold, and if the weighted sum or average is greater than or equal to the predetermined threshold, the given backbone link is determined to be significant. Otherwise, the given backbone link is determined to be not significant or to be non-significant.

[0082] In a third embodiment, the contributions of the backbone links 110.1-110.5 in several given situations are taken into account, for example in the three aforementioned situations, and the minimum contribution criterion is that a backbone link is significant if, in at least one of the given situations, the contribution of the backbone link is greater than a predetermined threshold. Alternatively, the minimum contribution criterion may be that a backbone link is significant if, in all the given situations, the contributions of the backbone link are greater than a predetermined threshold, or respective predetermined thresholds. Thus, for the given situations, the respective contributions of a given backbone link in these situations are compared to the same threshold, or respective thresholds (a threshold may be predetermined for each situation), and the given backbone link is determined to be significant if, in at least one situation or alternatively in all the situations, the contribution is greater than the predetermined threshold.Otherwise, the given backbone link is determined to be insignificant or to be non-significant.

[0083] For example, according to the third embodiment, the following three thresholds may be predetermined: .

[0084] - a first threshold for the nominal situation; - a second threshold for the peak situation; and - a third threshold for worst-case failure situations.

[0085] The first, second and third thresholds may have the same value or distinct values. A triplet of contributions is determined by the device 400 for any link 110.1-110.5. Thus, for a given backbone link, the first contribution of the triplet (nominal situation) is compared to the first threshold, the second contribution (peak situation) is compared to the second threshold and the third contribution (worst case failure situation) is compared to the third threshold. According to a first minimum contribution criterion, the given backbone link is significant if at least one contribution among the first, second and third contributions is greater than or equal to its respective threshold. According to a second alternative minimum contribution criterion, the given backbone link is significant if all of the first, second and third contributions are greater than or equal to their respective thresholds.

[0086] Other embodiments of step 205 allowing the determination of significant backbone links can be implemented according to the invention, which is in no way restricted to the three embodiments described above.

[0087] In the following, it is considered, for illustrative purposes only, that the backbone links 110.2, 110.3 and 110.5 are significant, and that the backbone links 110.1 and 110.4 are not significant.

[0088] At a step 206, the device 400 identifies a weakest backbone link among the significant backbone links 110.1, 110.3 and 110.4, the weakest backbone link being the backbone link for which a maximum additional bandwidth value that can be added to the first point of presence 101.1 without saturating the backbone link in its worst case failure is less than all the maximum additional bandwidth values ​​that can be added on the first point of presence without saturating the other significant backbone links respectively in the respective worst case failures of the other significant backbone links.

[0089] The maximum additional bandwidth value that can be added to the first point of presence 101.1 without saturating a given significant backbone link, can be determined by the device 400 from: - of a given significant backbone link capacity; - the load value of the given significant backbone link in the worst case failure for the given significant backbone link; - the contribution of the significant backbone link in the worst case failure for the given significant backbone link.

[0090] In particular, the maximum additional bandwidth value that can be added to the first point of presence 101.1 without saturating the given significant backbone link can be noted MAXBWj, j being an index of the given significant backbone link, varying between 1 and M, M being the number of significant backbone links which is greater than or equal to 2. [°° 91 1 MAXBWj J WCcontnb, j

[0092] in which LinkSizej is the capacity of the significant backbone link of index j, WCloadj is the load value in the worst case of failure for the significant backbone link of index j, and WCcontribj is the contribution of the significant backbone link of index j for the first point of presence 101.1 in the worst case of failure for the significant backbone link of index j.

[0093] Note that two differentiated values ​​of MAXBWj for the incoming traffic to the first point of presence 101.1 and the outgoing traffic from the first point of presence 101.1, can be determined by the device 400. In this case, MAXBWjjN represents the additional bandwidth value available in incoming traffic to the first point of presence 101.1 and saturating the significant backbone link of index j, and MAXBWjjOuT represents the additional bandwidth value available in outgoing traffic from the first point of presence 101.1 and saturating the significant backbone link of index j.

[0094] In this case, the capacity of the backbone link, the load value in the worst case of failure and the contribution in the worst case of failure are values ​​for the traffic entering the first point of presence 101.1, for the determination of MAXBWjjlN, and are values ​​for the traffic leaving the first point of presence for the determination of MAXBWj,out- Thus, the contributions determined in step 205 for the determination of the significant backbone links can be overall contributions for the traffic entering and for the traffic leaving the point of presence 101.1, while contributions differentiated for outgoing traffic and for incoming traffic can be considered during step 206 for the identification of the weakest backbone link

[0095] Alternatively, both the contributions determined in step 205 for determining the significant and non-significant backbone links, and the contributions considered in step 206 for identifying the weakest backbone link, are differentiated for the incoming traffic and for the outgoing traffic.

[0096] The weakest backbone link is thus the significant backbone link with index jMIN for which the MAXBWjM1N value is minimal (for which one of the two MAXBWjM1NjlNet MAXBWjM1NjOuTest values ​​is minimal), i.e. less than all the MAXBWj values, for j varying between 1 and M and j different from jMIN.

[0097] Thus, following the identification of the weakest significant backbone link jMIN in step 206, the device 400 obtains the value MAXBWjM1N which is therefore the additional bandwidth value which can be added to the first point of presence 101.1, before the weakest backbone link is saturated. This is therefore more generally the additional bandwidth value available for the first point of presence 101.1: this means that it is preferable not to grant new clients on the first point of presence more than MAXBWjmin, because otherwise, saturation of the weakest link risks being caused in the worst load case, and the network would thus lack resilience in the face of failure cases.

[0098] [Fig.3] is a graph 300 comprising: - a first axis 301 representing the indices i varying between 1 and iMAX, iMAX being equal to the number of backbone links connecting a given point of presence to the network 100. Thus, in the example of the first point of presence of [Fig.l], iMAX is equal to 5 for the first point of presence 101.1; - a second axis 302 representing the MAXBW value for all backbone links connecting a given point of presence to the network 100, only some of which are significant backbone links, as explained below.

[0099] Each horizontal position on the axis 301 corresponds to a given backbone link: for this given backbone link, at least one circle 310 is positioned along the axis 302 as a function of the MAXBV value; of the given backbone link of index i. When the incoming traffic and the outgoing traffic are not differentiated, only one circle is positioned on each horizontal position along the axis 301. When the incoming traffic and the outgoing traffic are differentiated, two circles can be positioned on a horizontal position along the axis 301, if the values ​​MAXBVijlN and MAXBVijOuT are distinct, which is most often the case in practice.

[0100] At a first horizontal position corresponding to the first backbone link 110.1, a first circle 311.1 is positioned vertically to represent the MAXBViiN value and a second circle 311.2 is positioned vertically to represent the MAXBViout- value Thus, according to graph 300, it is the outgoing traffic of the first point of presence 101.1 which is most constrained by the first backbone link 110.1, although the MAXBVij1N and MAXBVi values ​​are both high, compared to the other backbone links 110.2 to 110.5 as described in the following. In practice, different colors can be used to facilitate the differentiation between the MAXBVij1N value represented by the first circle 311.1 and the MAXBVi out- value represented by the second circle 311.2.

[0101] At a second horizontal position corresponding to the second backbone link 110.2, a circle 312 is positioned vertically to represent MAXBV21N and MAXBV2jout (in this case, the MAXBV21N value for traffic entering the first point of presence 101.1 is equal to the MAXBV2jOuT value for traffic leaving the first point of presence 101.1).

[0102] At a third horizontal position corresponding to the third backbone link 110.3, a circle 313 is positioned vertically to represent MAXBV3jin and MAXBV3j out (in this case, the MAXBV3jiN value for traffic entering the first point of presence 101.1 is equal to the MAXBV3jOuT value for traffic leaving the first point of presence 101.1).

[0103] At a fourth horizontal position corresponding to the fourth backbone link 110.4, a circle 314 is positioned vertically to represent MAXBV4jiN and MAXBViout (in this case, the MAXBV4jiN value for traffic entering the first point of presence 101.1 is equal to the MAXBV4jOuT value for traffic leaving the first point of presence 101.1).

[0104] At a fifth horizontal position corresponding to the fifth backbone link 110.5, a first circle 315.1 is positioned vertically to represent the MAXBV5jlNet value and a second circle 315.2 is positioned vertically to represent the MAXBV5 0ut- value. Thus, according to graph 300, it is the outgoing traffic of the first point of presence 101.1 which is most constrained by the fifth backbone link 110.5.

[0105] On graph 300, the size of each of the circles 311.1, 311.2, 312, 313, 314, 315.1 and 315.2 is proportional to the contribution of the corresponding backbone link 110.1-110.5 (for incoming and outgoing traffic) in one of the three situations described above, for example for the worst-case failure situation. It is for this reason that the smallest circles, namely circles 311.1, 311.2 and 314 correspond to the first and fourth backbone links 310.1 and 310.4, which are the backbone links determined as insignificant by the device 400 during the step 205 described previously.On the contrary, the circles 313, 313, 315.1 and 315.2, are of larger sizes, and correspond to the second, third and fifth backbone links 110.2, 110.3 and 110.5, determined as significant in step 205 by the device 400.

[0106] As shown in graph 300, the backbone link, among the significant backbone links 110.2, 110.3 and 110.5, having the smallest MAXBV value, i.e. having the circle in the lowest vertical position on graph 300, is the second backbone link 110.2 corresponding to circle 312. The second backbone link 110.2 is therefore the weakest backbone link of the first point of presence 110.1. Graph 300 also makes it possible to determine the second weakest backbone link, which in this case is the fifth backbone link 110.5, for the outgoing traffic of the first point of presence.

[0107] Without the selection of the significant backbone links of step 205, the fourth backbone link 110.4 would have been considered as the weakest backbone link, the values ​​MAXBV4j1N and MAXBV4jOut being lower than the values ​​MAXBV21N and MAXBV2jOut of the second backbone link 110.2 (the circle 314 is positioned vertically below the circle 312). However, since the fourth backbone link 110.4 is not significant, it is the second backbone line 110.2 which is advantageously retained as the weakest backbone link of the first point of presence 101.1.

[0108] The graph 300 can be displayed on a screen of the device 400, during step 206 for example. As a variant of the graph 300, a message can be displayed on a screen, identifying the weakest backbone link, and optionally the corresponding additional bandwidth value.

[0109] Referring again to [Fig.2], at a step 207, the additional bandwidth value available for the first point of presence 101.1 determined at step 206 is compared to a minimum available bandwidth criterion which is predetermined, and, depending on the result of the comparison, the device 400 generates or not an alert at a step 208. In the example considered above, the additional bandwidth value available for the first point of presence 101.1 is equal to MAXBV21N and MAXBV2jOut-

[0110] The minimum available bandwidth criterion may be a minimum available bandwidth threshold. The device 400 generates an alert at a step 208 if the additional bandwidth value available for the first point of presence 101.1 is less than the minimum available bandwidth threshold. No restriction is attached to the value of the minimum available bandwidth threshold.

[0111] The alert generated in step 208 can be displayed on a screen of the device 400, and / or can be transmitted to a remote entity, capable of communicating the generated alert to the operator in charge of planning the capacities of the network, or of a part of the network.

[0112] The invention thus makes it possible to present an alert in a synthetic and relevant manner to an operator in charge of planning network capacities. Such planning is thus facilitated, by avoiding issuing too many alerts, including included for backbone links which are not significant because they contribute little to the transport of traffic to / from the first point of presence 101.

[0113] Following the generation of the alert in step 208, or following step 207 if no alert is generated by the device 400, the method returns to step 201, in order to determine whether another point of presence for which the device 400 is in charge, has not been the subject of steps 203 to 210, in particular in order to determine the weakest backbone link and the additional bandwidth value available for the other point of presence.

[0114] In the present case, after the application of steps 203 to 210 to the first point of presence 101.1, described previously, steps 203 to 210 can be iterated for the third point of presence 101.3 of which the device 400 is also in charge.

[0115] Iterating the process for multiple points of presence makes it possible to improve capacity planning across the entire network, step by step.

[0116] According to embodiments of the invention, following step 205, the device 400 determines at a step 209, when several iterations of the method have been implemented for several points of presence, whether a given backbone link of the network 100 is determined to be insignificant for the two points of presence that it connects. If this is the case, the device identifies the given backbone link as 'not useful' at step 209, and can generate a backbone link alert 'not useful to the network' identifying the given backbone link, at a step 210. Otherwise, the method returns to step 201 and / or continues with step 206 for the current iteration of the method.

[0117] For example, the fourth backbone link 110.4 was determined to be insignificant for the first point of presence 101.1 during the first iteration of step 205. If at a subsequent iteration, applied to the fifth point of presence 101.5, the device 400 determines that the fourth backbone link 110.4 is insignificant for the fifth point of presence 101.5, then the device 400 identifies the fourth backbone link 110.4 as 'not useful'.

[0118] Such identification makes it possible to optimize the network. In particular, it makes it possible to differentiate: - cases in which a backbone link is significant for one point of presence located at its end, and insignificant for the other point of presence located at the other end. Such cases are common in so-called cascaded topologies when traffic is concentrated towards certain points of presence. These cases do not merit special attention from the operator;

[0119] - cases in which a backbone link is not significant for both points of presence that it connects, which indicates a case of 'non-usefulness' of the backbone link, which then deserves the attention of the operator to confirm that it is a possible optimization.

[0120] [Fig.4] illustrates the structure of the network planning assistance device 400, according to embodiments of the invention.

[0121] The device 400 comprises a processor 401 configured to communicate unidirectionally or bidirectionally, via one or more buses or via a direct wired connection, with a memory 402 such as a memory of the “Random Access Memory” type, RAM, or a memory of the “Read Only Memory” type, ROM, or any other type of memory (Flash, EEPROM, etc.). Alternatively, the memory 402 comprises several memories of the aforementioned types.

[0122] The memory 402 comprises at least one non-volatile memory in which the data used and / or resulting from the implementation of the steps of the method according to the invention described with reference to [Fig.2] are stored, temporarily or permanently.

[0123] In particular, the memory 402 can store: - the respective capacities of the backbone links of the network 100, or of the part of the network 100 for which the device is responsible for capacity planning; - the data obtained during step 203 for an iteration of the method for a given point of presence, in particular the average and peak load values ​​for all the backbone links connecting the given point of presence to the network 100; - the simulated data obtained during step 204 for an iteration of the method for a given point of presence, in particular the load value in the worst case of failure and the identifier of the backbone link causing the worst case of failure, for each backbone link connecting the given point of presence to the network 100; - the predetermined threshold(s) used for comparison with the contribution(s) for each backbone link, during step 205 described previously; - the minimum available bandwidth criterion applied during step 207 described previously.

[0124] The processor 401 is capable of executing instructions, stored in the memory 402, for implementing the steps of the method according to the invention, described with reference to [Fig.2],

[0125] The device 400 may further comprise a screen 403, for displaying the alert of step 208 described previously and / or the non-significant backbone link alert of step 210 described previously, and optionally for displaying the graph 300 shown in [Fig.3].

[0126] The device 400 comprises a first interface 403 configured to obtain the data from step 203, in particular the peak and average load values ​​of the backbone links connecting a given point of presence to the network 100, from an external entity.

[0127] The device 400 further comprises a second interface 404 configured to communicate with a network simulator to obtain the data from step 204 described previously.

[0128] The device 400 further comprises a third interface 405, configured to transmit the alert from step 208 and / or the backbone link alert not useful to the network from step 210, to an external entity.

Claims

Claims

1. Method for assisting in the planning of a network (100) comprising a plurality of points of presence (101.1-101.8) connected to each other in the network by links, called backbone links (110.1-110.5; 111.1-111.7), the method comprising the following steps, implemented for at least one of the points of presence (101.1): - determination (205) of significant backbone links, among the backbone links connecting said at least one point of presence to the network, as a function of respective contributions of the backbone links to the transport of data from and / or to said at least one point of presence; - identification (206; 209) of a weakest backbone link among the significant backbone links, the weakest backbone link being the significant backbone link for which a maximum additional bandwidth value that can be added to said at least one point of presence without saturating the significant backbone link, is minimum compared to maximum additional bandwidth values ​​that can be added to said at least one point of presence without saturating the other significant backbone links respectively.

2. A method according to claim 1, wherein the maximum additional bandwidth values ​​are determined from respective capacities and load values ​​of the significant backbone links.

3. Method according to claim 1 or 2, further comprising, based on a comparison (207) between the maximum additional bandwidth value that can be added on said at least one point of presence without saturating the weakest backbone link, and a minimum available bandwidth criterion, generating (208) an alert identifying the weakest backbone link for said at least one point of presence.

4. Method according to one of claims 1 to 3, further comprising obtaining respective load values ​​of the backbone links (110.1-110.5) connecting said at least one point of presence (101.1) to the network (100), the contributions of the backbone links to the transport of data from and / or to said at least one point of presence being determined from the respective load values ​​of the backbone links (110.1-110.5) connecting said at least one point of presence (101.1) to the network (100).

5. Method according to claim 4, obtaining the respective load values ​​of the backbone links (110.1-110.5) connecting said at least one point of presence (101.1) to the network (100) comprises: - obtaining (203) an average load value over a given period for any backbone link among the backbone links connecting said at least one point of presence to the network, the average load values ​​being associated with a nominal situation; and / or - obtaining (203) a peak load value over a given period for any backbone link among the backbone links connecting said at least one point of presence to the network, the peak load values ​​being associated with a peak load situation; and / or - obtaining (204), from a network simulator, for any given backbone link among the backbone links connecting said at least one point of presence to the network, a load value associated with a worst-case failure situation in the network for the given backbone link.

6. A method according to claim 5, wherein, for any given backbone link (110.1-110.5) connecting said at least one point of presence (101.1) to the network (100), at least one contribution to the transport of data from and / or to said at least one point of presence is determined (205) in at least one situation among the nominal situation, the peak load situation and the worst case failure situation for the given backbone link.

7. The method of claim 6, wherein, for any given backbone link (110.1-110.5) connecting said at least one point of presence (101.1) to the network (100), a first contribution to the transport of data from and / or to said at least one point of presence is determined (205) for the nominal situation, a second contribution to the transport of data from and / or to said at least one point of presence is determined for the peak load situation, and a third contribution to the transport of data from and / or to said at least one point of presence is determined for the worst-case failure situation for the given backbone link, and wherein the given backbone link is determined (205) as significant based on the three determined contributions.

8. Method according to one of claims 5 to 7, wherein the weakest backbone link is the significant backbone link for which the maximum additional bandwidth value that can be added on said at least one point of presence (101.1) without saturating the significant backbone link in the worst case failure situation for the significant backbone link, is minimal compared to the bandwidth values maximum additional bandwidth that can be added to said at least one point of presence and respectively saturating the other significant backbone links in the respective worst-case failure situations of the other significant backbone links.

9. Method according to one of the preceding claims, wherein the steps (203-210) of the method are applied to at least one other point of presence (101.2-101.8) of the network (100).

10. The method of claim 9, wherein if a backbone link (110.1-110.5) is identified as insignificant for said at least one point of presence (101.1) and for another point of presence (101.2-101.8) connected to said at least one point of presence via said backbone link, the backbone link is identified as being a backbone link not useful for the network.

11. Device (400) for assisting in the planning of a network (100) comprising a plurality of points of presence (101.1-101.8) connected to each other in the network by links, called backbone links (110.1-110.5; 111.1-111.7), the device comprising a processor (401) configured, for at least one point of presence, to: - determine significant backbone links, among the backbone links connecting said at least one point of presence to the network, as a function of respective contributions of the backbone links to the transport of data from and / or to said at least one point of presence; - identify a weakest backbone link among the significant backbone links, the weakest backbone link being the significant backbone link for which a maximum additional bandwidth value that can be added to said at least one point of presence without saturating the significant backbone link, is minimum compared to maximum additional bandwidth values ​​that can be added to said at least one point of presence without saturating the other significant backbone links respectively.

12. A computer program capable of being implemented in a generation module as defined in claim 11, the program comprising code instructions which, when executed by a processor (401), performs the steps of the method defined in one of claims 1 to 10.

Images