METHOD FOR DATA FLOW DISTRIBUTION BETWEEN WIDESPREAD NETWORKS AND COMMUNICATION SYSTEMS
The method addresses load imbalances in communication systems by dynamically adjusting flow distribution based on network load measurements and machine learning, optimizing resource use and adherence to SLAs.
Patent Information
- Application Number
- FR2021010894
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-10-14
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-10-14
AI Technical Summary
Existing load distribution strategies in communication systems are static and fail to adapt to dynamic traffic patterns, leading to load imbalances and violations of Service Level Agreements (SLAs) due to independent, local decision-making by routers that do not consider future load impacts.
A method for distributing data flows across multiple networks involving measuring network loads over a reference period, calculating target and adjusted proportions using machine learning, and minimizing distribution divergence to optimize resource use and balance loads.
The method effectively balances network loads, avoids peaks, and optimizes resource utilization by predicting and adjusting flow distribution based on historical and real-time data, ensuring compliance with SLAs and efficient resource allocation.
Smart Images

Figure 00000015_0000 
Figure 00000016_0000 
Figure 00000017_0000
Abstract
Description
Title of the invention: METHOD FOR DISTRIBUTING DATA FLOWS BETWEEN WIDE NETWORKS AND COMMUNICATION SYSTEM
[0001] The present invention relates to the distribution of data streams between several separate networks of a communication system. The invention also relates to a computer program for distributing data streams between several separate networks of a communication system. The invention further provides a multi-network communication system.
[0002] In the field of communication networks, the distribution of traffic between two or more paths of the same network uses a homogeneous multi-path routing mechanism in routers, designated by the acronym ECMP corresponding to the English expression "Equal-cost multi-path routing". IP packets of the same flow, with the same source and the same destination, are routed on several paths of equal cost. Paths of the same cost are defined as such by the metrics of the routing algorithm: the number of hops, the maximum bandwidth, and other parameters. The traffic is distributed equally between the different paths. In addition, ECMP is implemented in all routing solutions.For example, these routing solutions can be of the OSPF type, whose acronym corresponds to the English expression "Open Shortest Path First", of the ISIS type, whose acronym corresponds to the English expression "Intermediate System to Intermediate System", of the EIGRP type, whose acronym corresponds to the English expression "Enhanced Interior Gateway Routing Protocol", or of the BGP type, whose acronym corresponds to the English expression "Border Gateway Protocol". However, the distribution strategy is fixed and very limited; it does not take into account the effective load of the paths and cannot adapt to the dynamics of the flows.
[0003] To enable richer load distribution strategies, non-homogeneous multi-path routing strategies, generally referred to as "non-equal-cost multi-path routing", have been proposed in the literature. Thus, the DW solution, whose acronym corresponds to the English expression "Downward Path Criterion", does not limit the choice of paths to paths of the same cost, but also allows paths of slightly higher cost. The goal of these strategies is to use more paths to distribute traffic in the network.
[0004] Load sharing between different, separate and distinct networks exists also through PBR type mechanisms whose acronym corresponds to the Anglo-Saxon expression "Policy Based Routing", or PfR type whose acronym corresponds to the Anglo-Saxon expression "Performance based Routing", and software routing type, generally known by the acronym "SDWAN". These mechanisms make it possible to direct flows on technical criteria such as available bandwidth, the quality of the link with satisfaction of the SLA, whose acronym corresponds to the Anglo-Saxon expression "Satisfaction Level Agreement". The technical criteria can also be operational: the flow routing can be based on the type of application or the user profile. However, these mechanisms are limited in their capacity to adapt to the variabilities linked to heterogeneous traffic profiles.
[0005] Flow distribution is a problem that is also addressed in the context of server farms, also known as "cloud data centers". In this case, it is a matter of distributing application requests entering a data center to distribute them across the different servers, virtual machines, and containers. The request distribution function is handled by a load balancer.
[0006] Most load balancers implement static rules. They associate a weight with each server, the distribution of weights corresponding to the distribution target of incoming requests on the servers. This distribution can be uniform, also referred to by the English expression "Round Robin"; or non-uniform, which is generally referred to by the acronym WCMP, corresponding to the English expression "Weighted Round Robin". The distribution can be carried out according to whether the servers have the same CPU capacities or not. These strategies are also static, they work on average, but create a systematic imbalance in the presence of flows of heterogeneous characteristics, as occurs in networks and data centers.
[0007] The least load strategy, commonly referred to as "Least Loaded", on the other hand, takes into account the load status of each server and assigns incoming requests to the least loaded server. This is the strategy used by Kubernetes and HAProxy solutions. This strategy works for a very short decision-making period. The longer this period, the less effective it becomes because it works blindly with respect to traffic trends.
[0008] The networks and methods mentioned above are static. They do not take into account the impact of a decision on the future load on the networks. This produces a load imbalance over time and therefore violations of the SLA objective which requires respecting the proportions between networks.
[0009] A method of distributing incoming flows between several networks is known, where the The least loaded network is chosen in order to pass the IP packets of the same flow through it. Such a method requires obtaining the loads of each network, then comparing these loads in order to identify the lowest.
[0010] The difficulty with this approach is that the decision taken on the flows does not translate directly into the load. Indeed, some sessions are very short while others are very long. The latter remain in the network for a long time, and thus create an additional random load that these strategies cannot take into account.
[0011] Furthermore, since the decisions are taken locally by each router, the decisions between the input routers are independent. This leads to resource contention and therefore sub-optimal use of the available resources.
[0012] The invention aims to optimize the use of a communication system comprising a set of networks.
[0013] To this end, the invention relates to a method for distributing data flows in a communication system comprising a plurality of networks separated from each other, and a set of switching equipment connected to the networks and adapted to direct an incoming flow to one of the networks; at least one or each switching equipment of said set being configured so as to direct all the data frames of the same incoming flow to another switching equipment of said set via the same network of the plurality of networks; the method comprising the step: measuring flows of the networks in order to calculate an effective load specific to each network over a reference period; remarkable in that the method further comprises the following steps: obtaining a target distribution of flows between the networks, the target distribution comprising target proportions associated with the networks;obtaining an incoming load value [3 in the networks over a period following the reference period; obtaining a persistent proportion a of flows propagating in the networks during both the reference period and the following period; estimating, with control means, a distribution setpoint comprising an adjusted proportion for each network, at least one or each adjusted proportion being a function of: one of the target proportions, one of the actual loads, the incoming load value [3, and the persistent proportion a; and being capable of minimizing a distribution divergence score with respect to the target distribution over the following period; distribution of incoming flows, according to the adjusted proportions, between the networks by the switching equipment. ;
[0014] Thanks to the method according to the invention, the proportions applied at the input of the networks make it possible to correct, to compensate, the imbalance generated by the data flows which persist from one period to another. Load peaks are avoided, as are network saturation situations. As a corollary, the adjusted proportions make it possible to better distribute data flows and better exploit the available resources of the communication system.
[0015] According to other advantageous aspects of the invention, the method comprises one or more of the following characteristics, taken in isolation or in all technically possible combinations:
[0016] - The step of obtaining the incoming load value [3 comprises a measurement of data flow entering the communication system via the switching equipment during the following period, and the step of obtaining the persistent proportion a comprises a measurement of flow propagating during both the reference period and at least the following period via the switching equipment. - In the step of obtaining the incoming load value [3 and in the step of obtaining the persistent proportion a, the incoming load value [3 and the persistent proportion a are established using a machine learning algorithm. - The communication system comprises network output interfaces connecting each switching equipment to each network independently; in the measurement step the flows are measured at said network output interfaces, said outgoing flows being measured by the switching equipment. - The distribution divergence score comprises a sum of differences, each difference being associated with one of the networks and comprising the deviation between one of the target proportions and a network load ratio over the following period. - The control means comprise a control module separate from the switching equipment, the method further comprises a step of communicating the adjusted proportions to the switching equipment. - The network set comprises a first network and a second network, the target proportions comprising a first target proportion for the first network and a second target proportion for the second network, the second target proportion being greater than or equal to twice the first target proportion.
[0017] The invention also relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the latter to implement the method according to the invention.
[0018] The invention also relates to a communication system comprising control means, a plurality of networks separated from each other, and a set of switching equipment connected to the networks and adapted to direct an incoming flow towards one of the networks; at least one or each switching equipment of said set being configured so as to direct all the data frames of the same flow of data entering another switching equipment of said set via the same network of the plurality of networks; the communication system being configured so as to execute the data flow distribution method according to the invention.
[0019] According to other advantageous aspects of the invention, the system comprises one or more of the following characteristics, taken individually or in all technically possible combinations:
[0020] - The control means are integrated into at least one or each switching equipment. - Each switching device is a router and each network is a wide area network.
[0021] These characteristics and advantages of the invention will appear more clearly on reading the description which follows, given solely by way of non-limiting example, and made with reference to the appended drawings, in which:
[0022] [Fig-1] [Fig.l] is a diagram of a communication system according to the invention;
[0023] [Fig.2] [Fig.2] is a simplified representation of the communication system of [Fig.l]; and
[0024] [Fig.3] [Fig.3] is a diagram of a flow distribution method according to the invention.
[0025] [Fig.l] represents a communication system 2 according to the invention. The communication system 2 is capable of communicating any type of digital data.
[0026] The communication system 2 comprises a plurality of networks 4, 6, including a first network 4 and a second network 6. These networks 4, 6 are separate from each other. They are independent of each other. Each network 4, 6 comprises its own resources. The networks comprise its own nodes. The networks comprise its own segments, which connect their own nodes. The networks 4, 6 are wide area networks, generally designated by the acronym WAN which corresponds to the English expression “Wide Area Network”. They are wired networks, and / or networks which communicate by electromagnetic waves. On the other hand, the communication system 2 is a satellite communication system. The networks are of different types or not. They comprise different or identical maximum speeds.In the present example only two networks 4, 6 are shown, however the invention also applies to a telecommunications system 2 with more networks.
[0027] The communication system 2 further comprises a set of switching equipment 8 using the two networks. The switching equipment 8 is connected to the networks 4, 6. Each switching equipment 8 forms both an input switching equipment and an output switching equipment for the two networks, depending on whether a data stream enters or leaves one of the networks 4, 6 via the equipment 8. The switching equipment is bidirectional.
[0028] The switching equipment 8 is configured to select one of the networks 4, 6, and to switch an incoming flow 10 there. Furthermore, it is configured to direct all the data frames of the same incoming flow 10 to the same output switching equipment 8 via a single network of the plurality of networks. This makes it possible to improve the consistency of the application flows.
[0029] The communication system 2 comprises, or connects, external stations 12. These external stations 12 comprise at least one user interface and at least one server which are connected to different switching equipment 8. The server is formed by a server farm communicating with user interfaces via the communication system 2.
[0030] Each data frame, also called an IP packet, advantageously comprises a header with a source address and a destination address. Each data frame comprises bytes on which the data is recorded.
[0031] The communication system 2 comprises interfaces 11 between each network 4, 6 and each switching equipment 8. They connect each network to each switching equipment. The interfaces 11 are bidirectional. They each comprise, for each network and for each switching equipment, a network input interface and a network output interface.
[0032] All the IP packets belonging to the same incoming flow 10, for example a TCP session whose acronym corresponds to the English expression “Transmission Control Protocol”; or a UDP session whose acronym corresponds to the English expression “User Datagram Protocol”; are redirected to the same network input interface 11; and therefore to the same network. The communication system 2 makes it possible to transfer an input data flow 10 from a switching equipment 8 on the source side to a switching equipment 8 on the destination side. The switching equipment 8 comprises telemetry means, in particular in order to independently measure the flows at the output of the networks.
[0033] Each switching equipment 8 comprises means for measuring the flow entering each network over predetermined periods of time.
[0034] Similarly, each piece of equipment includes means for measuring outgoing flows.
[0035] For example, the flow is characterized by the average incoming flow rate during the period of time considered.
[0036] Furthermore, each device comprises means for storing for each time period an identifier of each incoming and outgoing frame and its size formed for example by the number of bytes constituting it.
[0037] The switching equipment 8 forms a border equipment. In the present embodiment, the set comprises six switching equipment 8. The equipment 8 are switching devices, switching to one or other of the networks. The devices are routers. Each router is configured to forward IP packets, or flow frames, from an input interface 11 of one of the networks to another output interface 11 of the same network. At least one or each router includes an input port, an output port, and a routing processor.
[0038] The communication system 2 comprises a control module 14 for the switching equipment. The control module 14 is functionally connected to each switching equipment 8. Each switching module 8 is capable of transmitting, at the end of each time period, the measurement of the incoming flow, the identifier of each frame entered on a network, its size and the network used, as well as the outgoing flow, the identifier of each outgoing frame, its size and the network of origin. The control module 14 centralizes the quantities of bytes measured at the output of each network. The flow measurements at the network output interfaces 11 make it possible to characterize the load in each network 4, 6 since these flows depend on the upstream network overloads. The control module 14 comprises a data storage unit, and a data processing unit such as a processor. The storage unit stores the measurements of the network flows.Thanks to the connections between the switching equipment 8 and the control module 14, the latter is able to communicate a distribution instruction to the switching equipment 8, so that they direct input flows 10 according to this distribution instruction. To do this, the control module 14 is configured so as to calculate an adjusted proportion dedicated to each network.
[0039] For this purpose, each piece of equipment comprises means for distributing flow according to the added proportion U4 and U6 calculated for each piece of switching equipment and transmitted by the control module 14.
[0040] [Fig.2] represents in a simplified manner the communication system 2 of [Fig.l], and illustrates the process of selecting a network from those possible via the networks 4, 6. Only two switching equipment 8 are represented for the sake of clarity.
[0041] In the present case, the target flow distribution, or target distribution, comprises a first target proportion p4, equal to one third, which applies to the first network 4; and a second target proportion p6, equal to two thirds, which applies to the second network 6. In the present case the second target proportion p6 is equal to twice the first target proportion p4. The sum of the target proportions is equal to 1. The target proportions are fixed a priori during configuration, for example according to the capacities of the networks.
[0042] The control module 14 is configured so as to communicate the same adjusted proportions U4 and U6 to each switching equipment 8; their sum is equal to 1. first adjusted proportion U4 corresponds to the proportion of flow entering 10 into an input switching equipment and switched into the first network 4, and the second adjusted proportion U6 corresponds to the proportion of flow entering 10 into the input switching equipment and switched into the second network 6. The control module is adapted to smooth, to balance the divergences from the target distribution, and this by calculating then applying an adjusted proportion, or distribution ratio; to each network.
[0043] [Fig. 3] represents a diagram of a method for distributing data flows in a communication system. The communication system corresponds to those described in relation to one of Figures 1 to 2.
[0044] The method comprises a step 100 of obtaining a target flow distribution p comprising target proportions associated with each network. The target proportions are defined as a function of optimal operating conditions of the communication system. At this step, the target proportions p4, p6 are communicated by the control module to the switching equipment. The target proportions are predefined or calculated following a modification of the operation of the communication system. When the maximum flow rate of a network decreases due to a malfunction, the target proportion corresponding to this network decreases and the target proportion of the other network increases.
[0045] The method comprises a step 102 of measuring network flows, optionally at the input and / or output of the networks, in order to calculate an effective load specific to each network over a reference period t, also called the first period. A load corresponds to a quantity of bytes leaving a network during the reference period t. It is calculated by the control module 14 from the quantities of outgoing bytes received from each switching equipment. In order to determine the specific loads during the reference period, the switching equipment only counts the bytes leaving the networks. Each switching equipment communicates these numbers of bytes for each network to the control module. The control module adds the numbers of bytes over the reference period t for each network and obtains the specific load of each network over the reference period.In the network flow measurement step 102, for the reference period t, we denote by X(t) the load on the first network, and Y(t) the load on the second network. The total load on all the networks is denoted by Z(t), and verifies the equation Z(t) = X(t) + Y(t). In the measurement step 102, the flows are measured at the network output interfaces, said outgoing flows being preferentially measured by the switching equipment.
[0046] The method comprises a step 104 of obtaining an incoming load value [3 in the networks during a period t+1 which follows the reference period t. The incoming load [3 corresponds to the number of bytes entering the communication system during the following period t+1. During the following period t+1, the switching equipment 8 observes the headers of the frames so as to identify the frames corresponding to incoming flows starting during the following period t+1. When a switching equipment 8 identifies such a flow, it counts the number of bytes. Each switching equipment 8 carries out this counting for each network over the following period t+1. Then, the switching equipment 8 transmits the number of bytes of each network to the control module which adds them, and obtains the incoming load [3. The following period and the reference period have the same duration. The following period is indexed t+1, and the loads on the networks are written X(t+1) and Y(t+1) over this following period.
[0047] The method comprises a step 106 of obtaining a persistent proportion a of flows propagating in the networks during the reference period t and the following period t+1. At the reference period t, the flow is broken down into a part PI ending during the reference period t, and a part P2 propagating both during the reference period t and the following period t+1.
[0048] The persistent proportion a is given by the formula a = P2 / (P1+P2). The PI part and the P2 part are identified using the frame headers and the number of bytes in the frames. During the reference period t and the following period t+1, the switching equipment identifies frames of the flows which end during the reference period t. They propagate in the reference period t but not in the following period t+1. The switching equipment 8 then transmits the number of bytes of these latter frames to the control module, which by adding them for all the networks, obtains the PL part. In addition, the switching equipment identifies and the flow frames propagating both during the reference period t and the following period t+1. The switching equipment 8 transmits the number of bytes of these latter frames to the control module.By adding these numbers of bytes for all networks, the control module obtains part P2. The persistent proportion a includes a combination of flows whose propagation time is greater than the reference period; and flows whose propagation time is less than said period but whose routing straddles the two periods t and t+1.
[0049] Thanks to the information collected, over the following period t+1, the estimated loads of the first network and the second network are written respectively X(t+1) = a * X(t) + [3 * U4 ; and Y(t+1) = a * Y(t) + [3 * U6 ; where U4 and U6 form unknowns to be identified. These equations form a predictive model of the load in the networks over the following period t+1. Similarly, we write Z(t+1) = a * Z(t) + [3.
[0050] The step 104 of obtaining the incoming load value [3 comprises the measurement of data flow entering the communication system via the switching equipment during the following period t+1, and the step 106 of obtaining the persistent proportion a includes a flow measurement propagating during the reference period t and at least the following period t+1. The incoming load value [3 and the persistent proportion a are a posteriori data, obtained after observations and calculations. Communication according to the adjusted proportions takes place after the following period, i.e. during a third period t+2. Thanks to the measurement steps on real and observed data, the distribution instruction is more precise.
[0051] According to an alternative of the invention, in the step 104 of obtaining the incoming load value [3 and in the step 106 of obtaining the persistent proportion a, the incoming load value [3 and the persistent proportion a are recovered or calculated using a machine learning algorithm, generally referred to as “Machine Learning”. According to another variant, the incoming load value [3 and the persistent proportion a are calculated on the basis of a traffic matrix. The traffic matrix describes the past exchanges of frames between the switching equipment 8 connected to the network 4, 6 considered.
[0052] In the context of a machine learning algorithm, the input data includes the parameters chosen from: the switching equipment(s) concerned, the date, the time, the source address, the destination address, frame header data, at least one of the network loads, at least one of the target proportions. The algorithm is trained using supervised learning. The training data comes from measurements, or from communication system simulations. The incoming load value [3 and the persistent proportion a are a priori data, which makes it possible to apply the distribution instruction from the following period t+1. This alternative makes it possible to accelerate flow control.
[0053] The method comprises a step 108 of estimation, by the control module 14, of the distribution setpoint comprising an adjusted proportion U4 or U6 for each network. In the estimation step 108, at least one or each adjusted proportion of a network depends on: the target proportion p4 or p6 of the network, the effective load X(t) or Y(t) associated with the network, the incoming load value [3, and the persistent proportion a; and being capable of minimizing, over the following period; a distribution divergence score with respect to the target distribution. The distribution divergence score is also called a balancing criterion.
[0054] The divergence score comprises a sum of differences; each difference being associated with one of the networks and comprising the deviation between its target proportion and a ratio of network load to the total load of the communication system 2 over a set T of periods. The sum is carried out over all the networks. The sum is a sum of absolute values of the differences, or a quadratic sum of said differences. Other mathematical formulas are considered.
[0055] Estimated load ratios over the following period are written in the form X(t+1 ) / Z(t+1 ) and Y(t+1) / Z(t+1) for the first and second network respectively. For each network, the estimated load ratio is equal to the estimated load of said network divided by the total load in the system, these loads being calculated over the following period t+1.
[0056] In the quadratic version, the divergence score J is written in the form:
[0057] l=y,r S / X(t + n) \ 2 +iF(t + n) V) » = o^Z(f + n) ' p4 ) \Z(t + n) '
[0058] where calculations are carried out for different periods n: t, t+1, ..., t+n, ... T.
[0059] The solution where T=1 is optimal since it allows us to get as close as possible to the p4 and p6 target distributions. It also limits the computations required, allowing for rapid application of the distribution instruction. Networks receive their corrections more quickly, and input flows are distributed more quickly according to conditions that avoid peaks while avoiding underutilizing part of the system.
[0060] From then on, we identify: U4 = p4 + a / [3 * (p4 * Y(t) - p6 * X(t)), and U6 = 1 - U4. These results optimize both the precision of the result and the required calculation time because they are based on a linear system of two equations with two unknowns.
[0061] When the system comprises at least three networks, the divergence score J is generalized for a system comprising more networks. A mathematical resolution makes it possible to identify adjusted proportions Ui for each ith network.
[0062] The method comprises a step 110 of communicating the adjusted proportions to the switching equipment during which the control module communicates the adjusted proportions to the switching equipment.
[0063] The method comprises a step 112 of distributing incoming flows, according to the adjusted proportions, between the networks by the switching equipment.
[0064] Following the distribution step 112, or between the communication step 110 and the distribution step 112, the method iterates. It again performs the network flow measurement step 102.
[0065] The reference period is a measurement period. It is shorter than the propagation time of at least one of the flows, preferably than the majority of the data flows. The measurement of the flows at the output of the networks is carried out throughout the measurement step. The duration of the reference period is between 1s and 2s.
[0066] According to one option of the invention, certain steps are carried out at the same time. For example, the obtaining steps are optionally carried out simultaneously.
Claims
1. Claims A method for distributing data flows in a communication system (2) comprising a plurality of networks (4, 6) separated from each other, and a set of switching equipment (8) connected to the networks and adapted to direct an incoming flow (10) to one of the networks (4, 6); at least one or each switching equipment (8) of said set being configured so as to direct all the data frames of a same incoming flow (10) to another switching equipment (8) of said set via a same network (4, 6) of the plurality of networks; the method comprising the step: • measurement (102) of flow of the networks (4, 6) in order to calculate an effective load specific to each network over a reference period; characterized in that the method further comprises the following steps: • obtaining (100) a target distribution of flows between the networks (4, 6), the target distribution comprising target proportions (p4, p6) associated with the networks; • obtaining (104) an incoming load value [3 in the networks (4, 6) over a period following the reference period; • obtaining (106) a persistent proportion a of flows propagating in the networks (4, 6) during both the reference period and the following period; • estimation (108), with control means (14), of a distribution instruction comprising an adjusted proportion (U4, U6) for each network (4, 6), • at least one or each adjusted proportion being a function of: one of the target proportions (p4, p6), one of the effective loads, the incoming load value [3, and the persistent proportion a; and being capable of minimizing a distribution divergence score with respect to the target distribution over the following period; • distribution (112) of incoming flow (10), according to the adjusted proportions, between the networks (4, 6) by the switching equipment (8).
2. A method of distributing data flows according to claim 1, characterized in that the step of obtaining (104) the incoming load value [3 comprises a measurement of data flow entering the communication system (2) via the switching equipment (8) during the following period, and the step of obtaining (106) the persistent proportion a comprises a measurement of flow propagating during both the reference period and at least the following period via the switching equipment (8).
3. A method for distributing data flows according to claim 1, characterized in that in the step of obtaining (104) the incoming load value [3 and in the step of obtaining (106) the persistent proportion a, the incoming load value [3 and the persistent proportion a are established using a machine learning algorithm.
4. Method for distributing data flows according to one of claims 1 to 3, characterized in that the communication system (2) comprises network output interfaces (11) (4, 6) connecting each switching equipment (8) to each network independently; in the measuring step (102) the flows are measured at said network output interfaces (4, 6), said outgoing flows being measured by the switching equipment (8).
5. Method for distributing data flows according to one of claims 1 to 4, characterized in that the distribution divergence score comprises a sum of differences, each difference being associated with one of the networks (4, 6) and comprising the deviation between one of the target proportions (p4, p6) and a network load ratio over the following period.
6. Method for distributing data flows according to one of claims 1 to 5, characterized in that the control means comprise a control module (14) separate from the switching equipment (8), the method further comprises a step of communicating the adjusted proportions to the switching equipment (8).
7. Method for distributing data flows according to one of claims 1 to 6, characterized in that the set of networks (4, 6) comprises a first network (4) and a second network (6), the target proportions comprising a first target proportion (p4) for the first network (4) and a second target proportion (4, 6) for the second network (6), the second target proportion being greater than or equal to twice the first target proportion.
8. Computer program product comprising instructions which, when the program is executed by a computer, cause the latter to implement the data flow distribution method according to one of claims 1 to 7.
9. A communication system (2) comprising control means, a plurality of networks (4, 6) separated from each other, and a set of switching equipment (8) connected to the networks and adapted to direct an incoming flow (10) to one of the networks (4, 6); at least one or each switching equipment of said set being configured to direct all the data frames of a same incoming data flow to another switching equipment (8) of said set via a same network (4, 6) of the plurality of networks; the communication system being configured to execute the data flow distribution method according to one of claims 1 to 7.
10. Communication system (2) according to claim 9, characterized in that the control means are integrated in at least one or in each switching equipment.
11. Communication system (2) according to one of claims 9 to 10, characterized in that each switching equipment (8) is a router and each network (4, 6) is a wide area network.
