Method for optimizing routing in a communication network

Abstract

The present invention relates to a computer-implemented method for optimizing the routing of data traffic in a communication network (1) using a quantum conceptual processor (6). Among the possible communication paths between each origin node (o) and each destination node (d) of the captured traffic requests (5a, 5b), a set (p) of potentially short communication paths is specified. The edges (e) in the set of potentially short communication paths are assigned respective utilization capacity limits. A partial capacity utilization of the edges (e) is calculated based on the respective utilization capacity limits of the edges (e). The calculated partial capacity utilization is then formulated as a term of a quadratic stress function. Using the quantum conceptual processor (6), an optimized routing is determined by selecting for each traffic request (5a, 5b) one short communication path (p1, p2) from the set of potentially short communication paths (p) such that the quadratic stress function is minimized.

Description

Technical Field

[0001] The present invention relates to a computer-implemented method for optimizing the routing of data traffic in a communication network having a plurality of communication nodes connectable via edges (links) of a communication path for the routing of data traffic. The present invention also relates to a quantum concept processor configured to execute such a method and a computer program implemented to execute such a method.

Background Art

[0002] Today's requirements regarding data traffic in communication networks are increasing dramatically in these times. With the recent introduction of 5G, ever more devices and applications are pushing data traffic to new peaks. Furthermore, the increasing requirements for digitized and decentralized work and the increasing streaming requirements in the domestic environment of private households also contribute significantly to this trend. The increasing amount of data transferred via communication networks such as the Internet poses a major challenge for service providers. In order to avoid network congestion and degradation of the user experience, traffic engineering techniques are deployed to complement the relatively slow and expensive expansion of the network infrastructure.

[0003] The most widely deployed engineering techniques for data traffic management in communication networks operate on the premise of the shortest path calculated with respect to link weights. These weights are often related to link capacity, i.e., the maximum amount of data per unit time that can be routed through the link from the start node to the end nodes respectively connected by that link at each edge. The final routing of a data stream from the origin node to the destination node is based on the identified path of the shortest path found, taking into account intermediate nodes and a given link or edge. As a result, the simplest and most practical technique for guiding traffic demands is through the manipulation of these link weights, also known as link metrics or IGP (Interior Gateway Protocols) metrics. The higher the link weight of an edge, the higher the probability that data will be routed through each edge. Following this approach, in a reactive way, link weights are locally adapted whenever a particular link tends to be overloaded. In a more systematic way, this problem is further addressed by applying a linear integer computer program, where the optimization goal is to minimize the maximum link capacity usage in the network. However, the task of finding a globally optimal set of metrics is very complex. In terms of computational complexity, this task is NP-hard because each link metric can potentially affect a large number of communication paths.

[0004] The linear optimization techniques applied so far quickly reach their limits when considering actual non-linear conditions such as redundancy, geographical subgroups or subdomains (e.g., the European network and the US network considered in one model), inclusion of satellites, quality of service, QoS, relationships, etc. Furthermore, the known techniques often lead to problems of unused capacity utilization and link capacity overload in the communication paths within the network, where many links are close to their capacity limits. SUMMARY OF THE INVENTION

[0005] Accordingly, the problem of the present disclosure is to provide an enhanced technique that enables optimized utilization of communication paths within a communication network with respect to capacity limitations, thereby achieving optimized routing.

[0006] This problem is solved by the method according to claim 1. Further embodiments are described in the dependent claims and the following description.

[0007] This method is a computer procedure for optimizing the routing of data traffic in a communication network having a plurality of communication nodes. The communication nodes are connected by the edges of the communication network. A series of edges generates a communication path for the routing of data traffic. Thus, the edges of the communication path in this context describe the connection between two adjacent nodes within the communication path.

[0008] The method includes: - capturing a set of traffic requests, each traffic request specifying the transfer of a determined data volume from a source node to a destination node among the plurality of communication nodes; - specifying a set of communication paths that are likely to be short among the possible communication paths between each respective source node and each respective destination node specified in the set of traffic requests, with each edge in the set of communication paths that are likely to be short assigned a respective utilization capacity limit; - calculating, for the set of traffic requests, the partial capacity utilization rate of the edges in the set of communication paths that are likely to be short, the partial capacity utilization rate being calculated based on each respective utilization capacity limit; - formulating the calculated partial capacity utilization rate as a term of a quadratic stress function; - using a quantum theory concept processor to determine an optimized routing by selecting, for each traffic request in the set of traffic requests, one short communication path from the set of communication paths that are likely to be short such that the quadratic stress function is minimized.

[0009] This method reliably addresses the problem of routing network requests in a communication network along an optimized short path, thereby avoiding in an optimized way the exceeding of link capacities within the network.

[0010] By applying this method, for each given traffic request, one optimal option for a short communication path can be selected from a set of potentially short communication paths. This selection is made such that the capacity of all edges (links) within the communication paths used within the network is respected as an upper limit for the total volume of traffic requests routed along them.

[0011] A "traffic request" in this context is modeled as a 3-tuple, defining a source node (source of the data stream), a destination node or target node (destination of the data stream), and a determined data traffic to be transferred between the source and the destination. The focus is on providing a continuous data stream over the network, which is modeled and routed such that data is not lost during transmission by exceeding the specified capacity on a given transport link. The measurement of such a data stream request or the data transfer rate of the request is currently specified in Gbps (gigabits per second).

[0012] A "short" communication path in this context can be selected as a path having determined intermediate nodes between each respective source node and each respective destination node specified in a set of traffic requests, or as the directly shortest path. According to a detailed implementation further described below, a "short" communication path having determined intermediate nodes (so-called segment nodes) between each respective source node and each respective destination node is specified in a set of traffic requests, each consisting of the shortest path from the source to the segment node and the shortest path from the segment node to the destination node.

[0013] By selecting such short paths between each source node and each destination node, more or less direct routing can be achieved, thereby avoiding long communication paths that impose an adverse burden on a plurality of edges along each communication path with a rather high capacity utilization rate in an inefficient manner.

[0014] Among the possible communication paths between each source node and each destination node, a set of communication paths that may be short is determined in advance, for example, by applying any other algorithm among the Dijkstra algorithm or other algorithms of the class of efficient shortest path algorithms.

[0015] To address the complexity of the optimization problem described above, a quadratic optimization problem can be formulated by calculating the partial capacity utilization rate of the edges within the set of communication paths that may be short and formulating the calculated partial capacity utilization rate as a term of a quadratic stress function. The application of such a quadratic optimization problem has the effect of being able to formulate a quadratic stress function that greatly disadvantages a high capacity utilization rate on individual edges of the communication path.

[0016] In this way, using a quantum theory concept processor, an optimized routing is determined by selecting one short communication path from the set of communication paths that may be short for each traffic demand in the set of traffic demands such that the quadratic stress function is minimized. The minimum value of the quadratic stress function is preferably a global minimum, but can also be a local minimum.

[0017] Therefore, the method has the technical effects and advantages of achieving a uniform minimum utilization of the network with respect to capacity limitations and the distribution of the distance to the capacity limitations within the network. At the same time, the method provides the selection of short paths between each source node and each destination node for each traffic demand in the communication network.

[0018] As described above, the underlying quadratic optimization problem is very complex. This is not only due to the fact that one selected communication path can affect other communication paths and the huge amount of data traffic managed between multiple source nodes and destination nodes within the network. Also, this problem is very complex because there are many practical constraints that must be considered. As more constraints are implemented, such problems become more complex and difficult to solve. This is a problem or difficult when a traffic engineering solution is needed quickly, for example, in response to an unexpected network failure or when considering additional practical constraints such as latency (the shortest possible path, the minimum number of IP hops possible), redundancy (the model should be redundant against the failure, planned outage, or maintenance of one or more / many edges), domain (EU, US) or hierarchy (core network, access network). The method described herein advantageously demonstrates its strength compared to conventional approaches where the fundamental problem becomes increasingly complex. In other words, for complex optimization problems considering practical constraints as described above, the method described herein has a significant strength over the prior art.

[0019] The method described herein utilizes an approach inspired by quantum computing. The calculation of the optimal solution of a quadratic stress function for determining an optimized communication path for a set of traffic demands is performed by a so-called quantum theory concept processor. As a quantum concept processor in the context of the present disclosure, a so-called "Ising model" or a processor that solves an equivalent binary problem without quadratic constraints is defined. For example, a processor configured to solve an optimization problem using quantum annealing or quantum annealing emulation. Such a processor is based on, for example, conventional hardware technologies such as complementary metal oxide semiconductor (CMOS) technology. An example of such a quantum concept processor is Fujitsu's Digital Annealer. Alternatively, any other quantum processor can be used in the method described herein, and in the future, technologies based on actual qubit technology can also be used. Further examples of such quantum concept processors include not only D-Wave's quantum annealers (e.g., 5000Q), but also quantum gate computers (IBM, Rigetti, OpenSuperQ, IonQ, or Honeywell) that utilize quantum optimization algorithms such as QAOA or VQE.

[0020] In other words, the quantum concept processor defined herein is a processor that realizes the concept of minimizing a so-called binary optimization without quadratic constraints (QUBO) function, either by classical techniques of special processors, quantum gate computers, or quantum annealers.

[0021] In at least one implementation, the method comprises: - specifying a set of possible segment nodes among a plurality of communication nodes, each possible segment node defining a short possible communication path as a member of a set of short possible communication paths between a source node and a destination node and as an intermediate node; - formulating segment node terms in a secondary stress function, each segment node term connecting the calculated partial capacity utilization rate of the edges of its respective short possible communication path to those segment nodes within the set of segment nodes that may be connected to its respective short possible communication path; - using a quantum concept processor to calculate the segment node terms and select segment nodes within the set of possible segment nodes such that the secondary stress function is minimized for determining an optimized routing.

[0022] In this way, for each traffic requirement, one or more segment nodes are individually determined, and the so-called Segment Routing (SR) protocol can be implemented. Accordingly, a set of possible segment nodes among a plurality of communication nodes is specified, and each of the possible segment nodes defines a potentially short communication path as a member of a set of potentially short communication paths between a source node and a destination node. Segment Routing (SR) is a concept that defines alternative communication paths between individual source nodes and destination nodes. For each pair of a source node and a destination node, a sequence of one or more so-called segment nodes is selected, and then the path between the source and the destination is given as the concatenation of the shortest path between the source and the first segment node, the shortest path between the segment node and its respective subsequent node in the sequence of segment nodes, and the shortest path between the last segment node and the destination node. A specialization of SR is nSR (n-Segment-Routing), where a maximum of n - 1 segment nodes, and thus n shortest path segments, are possible as the route between the source and the destination. This is the so-called SPRING (Source Packet Routing in Networking) concept. Since SR can revert to a specific variation or adaptation in the routing of data traffic by using one or more segment nodes between each source node and each destination node, it provides additional degrees of freedom. This provides an elegant compromise between static shortest paths and flexible, variable routing of data traffic.

[0023] For example, in an exemplary case where the direct shortest path between each source node and each destination node is not available, such as due to construction work or maintenance within the network, the SR provides additional degrees of freedom. In this way, the selected detour can be selected in the step of determining the optimized path in the method described above, yet it is still an optimized path that is short in the sense described above. However, considering the overall optimized distributed capacity utilization rate within the network, the SR provides additional degrees of freedom. In this way, different short paths for different traffic demands (e.g., via different segment nodes) can be selected to avoid overloading or a significant increase in the capacity utilization rate at each edge of the communication path within the network.

[0024] It is possible to calculate the optimal solution (minimum value) of the quadratic stress function by connecting the calculated partial capacity utilization rate of the edges within the possible communication paths, taking into account the assignment of different segment nodes to the communication paths for different traffic demands according to the SR (in the segment node term). In this way, an optimized selection of one or more respective segment nodes for each traffic demand can be achieved to satisfy the optimization problem described above. Therefore, the impact of the communication path selected for one traffic demand on other possible communication paths for other traffic demands can be mitigated. Therefore, by applying the degrees of freedom of the SR, different short communication paths for different traffic demands can be directed to different segment nodes within the network to distribute the overall capacity utilization rate within the network. The optimized selection of each segment node is executed by a quantum concept processor.

[0025] The set of possible segment nodes can be preselected for each traffic request, considering that for each subset of all possible segment nodes, the corresponding communication path is selected such that it is close to the shortest path between each source node and each destination node with respect to a predetermined capacity utilization rate of the relevant edges within the communication path. A segment node can represent a router, i.e., a node or an interface, i.e., a physical network element such as a node.

[0026] In at least one implementation of the method, the segment node term is calculated considering the path condition that each traffic request of a set of traffic requests is routed along the shortest path or through exactly one segment node between each source node and each destination node. Such a path condition forms a constraint or "boundary" on the way to implement the so-called two-segment routing (2SR) protocol. In an actual network, 2SR already provides sufficient flexibility and freedom compared to general SR, yet it has been found that it avoids implementing an extremely large number of segment nodes. This keeps the implementation cost low. Additionally, 2SR enables optimized routing that is close to the shortest path for all traffic requests. For each traffic request of a set of traffic requests, the quantum concept processor selects either the shortest path from each source node to each destination node or a 2SR path across one selected segment node from the set of possible segment nodes, as described above, to achieve the minimum value of the quadratic stress function.

[0027] In an alternative implementation of this method, the segment node terms are calculated taking into account the routing condition that each traffic requirement of a set of traffic requirements is routed along the shortest path or via a plurality of segment nodes between each source node and each destination node. Such routing conditions form a constraint or "boundary" on the method for implementing the Segment Routing (SR) protocol instead of the 2-Segment Routing (2SR) protocol.

[0028] In at least one embodiment of this method, the segment node terms are calculated taking into account cost conditions such that the number of selected segment nodes is minimized. In this way, an additional optimization goal is formulated and considered in a method related to minimizing the total number of assigned segment nodes in order to minimize deployment and maintenance costs. In a theoretical approach, the more segment nodes are assigned, the more optimal the solution of the quadratic stress function found by the quantum theory concept processor becomes. However, a large number of segment nodes means high costs for segment node deployment and maintenance. Using the additional optimization goal of minimizing the total number of assigned segment nodes, minimization of the quadratic stress function can be achieved by the quantum theory concept processor, thereby achieving reduction and minimization of the total cost of the assigned segment nodes. Depending on which optimization to focus on, both optimization criteria can be balanced against each other.

[0029] In at least one implementation of this method, the quadratic stress function is formulated as a quadratic unconstrained binary optimization (QUBO) function. This QUBO function serves as the "input" to a quantum concept processor that solves this optimization problem for the optimized routing of all traffic demands, by the method described above. Generally speaking, a QUBO is a quadratic polynomial of binary variables represented within a quantum concept processor as bits or qubits (hereinafter, qbits). In the context of the optimization problem of the present disclosure, the QUBO function represents, as a function of different qbits, the sum of the possible contributions of the partial capacity utilization rates of each edge within a possible communication path, and each qbit represents a choice of a path alternative that can assume a value of "0" or a value of "1". To solve the quadratic optimization problem (quadratic stress function), the quantum theory concept processor executes different settings of different qbits to find a solution that minimizes the quadratic optimization problem. In this way, the QUBO representation of the optimization problem has elegant properties with respect to the quantum concept computing applied here.

[0030] In at least one implementation of this method, the quadratic stress function, and at least one of the path condition and the cost condition, are each weighted and combined towards a global QUBO function, as described above. This has the advantage that different partial optimization problems can be weighted relative to each other. For example, the quadratic stress function and the cost condition, which depend on the number of selected segment nodes, are weighted relative to each other according to the priority of cost reduction or uniform traffic distribution within the network.

[0031] In at least one implementation, the method further includes: - selecting a subset of traffic demands from a set of traffic demands; - executing the method for the subset of traffic demands; - storing the optimized routing determined for the subset of traffic demands; - updating, in consideration of the optimized routing determined for the subset of traffic demands, the remaining utilization capacity limits of each edge within a set of possible short communication paths.

[0032] In this way, a decomposition strategy can be followed. This is advantageous or even necessary to process the described method despite the limited hardware performance of the quantum conceptual processor for solving the quadratic optimization problem. Assume that the number of variables required to formulate the quadratic optimization problem is large. Today, the possibilities of current quantum conceptual processors are still limited. Therefore, very complex quadratic optimization problems have to be decomposed into several partial solutions that can be processed iteratively to find the optimal solution. In each iteration, one partial solution is found by the quantum conceptual processor.

[0033] For example, considering the QUBO formulation of the quadratic optimization problem formulated as an SR problem, as described above, the number of bit variables (Qubits) required to formulate the optimization problem is, for example, |D|x|S|, where |D| is the number of traffic demands to be processed and |S| is the number of segment nodes considered for each demand. In a typical dataset underlying the present disclosure, the number of traffic demands |D| is typically in the thousands, and the set of segment nodes |S| may include approximately 50 segment nodes.

[0034] Furthermore, considering that today's typical quantum conceptual processors can solve quadratic optimization problems on the order of 10,000 bit variables (Qubits), the overall optimization problem has to be solved iteratively with the help of problem decomposition.

[0035] Therefore, the optimization problem can be iteratively decomposed by the strategy described above, where a subset of traffic demands is selected from the set of traffic demands and the described method is executed for the subset of traffic demands. Subsequently, the determined and optimized routing for the subset of traffic demands is stored, and considering the determined and optimized routing for the subset of traffic demands, the remaining utilization capacity limits of each of the edges within the set of possible short communication paths are updated.

[0036] According to an exemplary implementation, to decompose the problem, traffic requests are sorted in descending order by volume and split into separate parts of the requests. In particular, requests with the largest volume can be considered first, and then requests with the smallest volume can be considered. Then, starting from the request with the largest volume, the corresponding routing optimization problem is iteratively formulated and optimized for each separate part of the request with the help of a quantum concept processor. The partial solutions for each separate part of the request are stored, and the edge capacity is reduced by the usage rate corresponding to the partial solution found in the previous iteration. Then, the method is processed with the next subset of requests. In this way, large-volume traffic requests that require a higher portion of the edge capacity are considered in the initial iterations of the algorithm, while smaller-volume requests are considered in later iterations of the algorithm and distributed over the remaining capacity of the network, so that a network with little capacity consumption can be accessed.

[0037] In at least one implementation, the method further includes: - selecting a subset of possible segment nodes from a set of possible segment nodes; - executing the method on the subset of possible segment nodes.

[0038] Such an implementation provides a more specialized decomposition strategy that takes into account SR, as described above. To further save variables (Q bits), the set of possible segment nodes |S| can be restricted to a subset of possible segment nodes from the set of possible segment nodes for each traffic request. Then, in the sense of the optimized solution (the minimum value of the quadratic optimization problem), the segment nodes calculated and selected for each traffic request are stored before the next iteration of the method. For example, the subset of possible segment nodes for each traffic request can be preselected as the segment nodes that are very close to the shortest path between the respective source node and destination node of each traffic request.

[0039] According to a further implementation, one or more of the decomposition procedures described above are repeatedly executed for the remaining traffic requests until all traffic requests of the set of traffic requests are processed.

[0040] The above problem is also solved by a quantum concept processor claimed in the appended claims. The quantum concept processor is configured to execute one or more steps of the method as described above. According to an exemplary implementation, the quantum concept processor is a digital annealing processing unit. This unit can be specially configured to perform quantum annealing or quantum annealing emulation as described above. The quantum concept processor can be of any type among those described above.

[0041] The above problem is also solved by a computer program comprising instructions that, when the program is executed by one or more processors, cause each of the one or more processors to execute one or more steps of the method described above. At least one of these processors is, for example, a quantum concept processor as described above. Other processors can be configured to process preliminary or iterative steps of the method as described above by executing the computer program.

[0042] Furthermore, in order to verify the optimized path determined by the method as described above, the above problem is also solved by a workplace for a network planner. Such a workplace has, for example, verification means configured for (automated or semi-automated) verification of the optimized path determined by the method as described above. This functions such that the network planner can verify the optimization result found by the method as described above. The verification means can be implemented in software and / or hardware. For example, the workplace can communicate with or be connected to a system including a quantum concept processor that executes the method as described above. Then, the results can be passed on to the workplace.

[0043] Furthermore, the above problem is also solved by an interface configuration including one or more interfaces to a plurality of communication nodes of a communication network through which data traffic is routed, the interface configuration being configured to automatically deploy the optimized routing determined by the method as described above to the communication nodes of the communication network. In this way, the optimized routing determined by the method as described above can be (automatically or semi-automatically) deployed to the plurality of communication nodes of each communication network. For example, the interface configuration can communicate with or be connected to the workplace as described above, or a system including a quantum concept processor that executes the method as described above. The results can be passed on to the interface configuration.

[0044] Furthermore, as a preliminary measure for one or more of the steps described above of the computer-implemented procedure, an interface for reading parameters from the communication network before each optimization and inputting such parameters into the described computer-implemented optimization procedure can be implemented or used. The parameters include, for example, network configuration, adjacency information for graph description of the network, available capacity within the network, and expected traffic demands.

[0045] Any aspect, feature, effect, and measure described alone or in combination in connection with the method described above applies, or similar expressions can be found, to the aspects, features, effects, and measures described alone or in combination in connection with the quantum concept processor or computer program described above, and vice versa. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present invention will be further described below in view of several implementations with the aid of multiple drawings.

[0047]

Figure 1

Figure 2A

Figure 2B

Figure 3

Figure 4A

Figure 4B

Figure 4C

Figure 4D

Figure 5

[0048] FIG. 1 shows an exemplary configuration of a communication network 1 having exemplary routing of traffic demands 5a and 5b according to a conventional approach. The communication network 1 includes a plurality of communication nodes 2, and a connection 4 between two adjacent communication nodes 2 is called an edge. This is shown between a communication node 2 and another communication node 2c, and they can communicate with each other via the connection 4. Depending on the historically grown configuration and implementation of the communication network 1, some communication nodes 2 are aggregated into so-called aggregation nodes 3. As exemplarily shown in FIG. 1, for example, the communication node 2a is aggregated into the aggregation node 3a, and other communication nodes 2b, 2d, 2e are aggregated into the aggregation node 3b.

[0049] The communication node 2 is, for example, a so-called label edge router (LER) for routing incoming and outgoing data traffic within the network 1. The aggregation node 3 is called a metanode and is an aggregation zone of LERs in a specific area of the network 1. For example, the aggregation node 3 is a concentrated aggregation zone of a determined economic region or city where communication is to be carried out. In other applications, the aggregation node 3 can be, for example, an entity such as an industrial network or a traffic network.

[0050] The communication network 1 is generally partially meshed. This means that not all communication nodes 2 are connected to all other communication nodes 2. Instead, there are only some connections 4 (see the dotted connections), for example, resulting from the historical development of the network 1, between some of the communication nodes 2 implemented in the network 1. The connection 4 between each communication node 2 is implemented, for example, by an optical fiber connection. However, other technologies such as wireless technologies (e.g., 5G) and copper wire / DSL technologies are also generally applicable.

[0051] As described above, FIG. 1 shows a particular scenario of traffic requests 5a and 5b, according to which a particular amount of data must be transferred between each communication node 2 within network 1. As illustratively shown, the first traffic request 5a is between communication node 2a within aggregation node 3a and another communication node 2f within aggregation node 3d. The second traffic request 5b is between communication node 2e within aggregation node 3b and, although repeating, communication node 2f within aggregation node 3d. Thus, each traffic request 5a and 5b defines a determined amount to be transferred from a source node to a destination node. In the exemplary scenario according to FIG. 1, the source node for traffic request 5a is communication node 2a, while the destination node for traffic request 5a is communication node 2f. Similarly, for traffic request 5b, the source node is communication node 2e and the destination node is communication node 2f.

[0052] In an alternative implementation, the traffic requests can be defined as requests between the aggregation nodes 3 regardless of which internal communication node 2 within each aggregation node 3 the communication starts or ends at. For example, requests 5a, 5b can be defined as a request between aggregation nodes 3a and 3d (request 5a), a request between aggregation nodes 3b and 3d (request 5b). In such an implementation, there is a "virtual" edge between each aggregation node and its internal communication node, and the virtual edge has a very high capacity. This leads to the effect that it does not matter at which internal communication node 2 within each aggregation node 3 the communication starts or ends.

[0053] Each of the traffic requests 5a and 5b imposes a burden on Network 1 in terms of the utilization rate of the network's capacity, that is, the utilization rate of the capacity of each connection 4 of the possible communication paths between each communication node 2 within Network 1. In the exemplary scenario of FIG. 1, traffic request 5a is transferred from communication node 2a to communication node 2f via communication nodes 2b, 2c, 2d, 2e, and 2f. In parallel with this, traffic request 5b is simply transferred via the connection 4 between communication nodes 2e and 2f. In this scenario, two drawbacks occur. The first drawback is that the communication path for transferring traffic request 5a is long and complex through Network 1. This transfer embeds multiple communication nodes 2 and connections 4 within Network 1 to transfer traffic request 5a. The second drawback lies in the fact that both traffic requests 5a and 5b are ultimately transferred via the connection 4 between communication nodes 2e and 2f. Therefore, the link capacity of the connection 4 between nodes 2e and 2f becomes a significant load. This can lead to an overload of the connection 4 between nodes 2e and 2f, resulting in an increase in latency or data loss, etc.

[0054] FIG. 2A shows an exemplary configuration of a communication network 1 with an exemplary routing of traffic requests 5a and 5b (refer to the above) according to an alternative approach. In the scenario according to FIG. 2A, an intermediate node s, hereinafter referred to as a segment node s, is determined to route traffic request 5a. The segment node s is configured within an aggregation node 3b together with other communication nodes 2b, 2e. Comparing with FIG. 1, for example, communication node 2d (refer to FIG. 1) is declared as a segment node s in the scenario of FIG. 2A. Thereby, starting from the repeating communication node 2a, traffic request 5a is transferred via an alternative communication path following communication nodes 2b, segment node s, communication nodes 2e, 2f.

[0055] The scenario according to FIG. 2A is such that the communication path for the transfer of traffic request 5a approaches the short path or the shortest path strategy, thereby keeping the number of communicating nodes 2 and connections 4 involved in network 1 low (at least lower than in the scenario of FIG. 1), which is more advantageous than the scenario of FIG. 1. However, even in the scenario of FIG. 2A, other drawbacks remain, according to which both traffic requests 5a and 5b pass through this connection of network 1 in the connection 4 between nodes 2e and 2f, and still a heavy load is imposed.

[0056] FIG. 2B shows an exemplary configuration of the communication network 1 according to FIGS. 1 and 2A, where the exemplary routing of traffic requests 5a and 5b follows the approach according to the invention. In the scenario of FIG. 2B, an optimized segment node s is selected, which is within the aggregation node 3e instead of the aggregation node 3b according to FIG. 2A. Thereby, starting from node 2a, traffic request 5a is transferred within the communication path to communication nodes 2b, 2c, segment node s, and communication nodes 2g and 2f. The other traffic request 5b is transferred between the two communication nodes 2e and 2f as in the scenarios according to FIGS. 1 and 2A.

[0057] Therefore, the scenario according to FIG. 2B transfers traffic request 5a on a still relatively short path between the source node 2a and the destination node 2f. However, the substantial value of the scenario according to FIG. 2B lies in the fact that the traffic request 5a on the final route segment towards its destination 2f is transferred via the connection 4 between nodes 2g and 2f instead of via the connection 4 between nodes 2e and 2f. Thereby, only the traffic of traffic request 5b is loaded on the connection 4 between node 2e and node 2f.

[0058] Therefore, the scenario of Figure 2B solves the drawbacks of the approaches according to Figures 1 and 2A, thereby achieving short-path communication within Network 1 with a uniform and optimized distribution of the overall capacity utilization rate of the connections 4 within Network 1 for all traffic demands 5 that must be transferred within Network 1.

[0059] In the following, the implementation of the approach according to Figure 2B will be described in more detail.

[0060] The optimization problem to be solved is to determine an optimized routing through Network 1 by selecting, for each traffic demand 5, one short communication path from a set of potentially short communication paths such that a mathematically formulated quadratic stress function (core optimization problem) is minimized. This, combined with the effect that the overall capacity utilization rate of the connections 4 within the selected communication paths can be uniformly minimized within Network 1 for all traffic demands 5 within Network 1, serves the purpose of selecting each communication path as close as possible to the shortest path strategy and has a technical effect. Thereby, it is avoided that some of the connections 4 in each communication path are subjected to a large or excessive load, while the small load on other connections 4 can significantly reduce such stress.

[0061] To achieve the above effects, a computer-implemented algorithm method for optimizing routing within the communication network 1 is implemented. This is described below.

[0062] Figure 3 shows a schematic diagram of possible communication paths p1 - p4 for routing traffic requests between source nodes o1, o2 and destination nodes d1, d2. In the exemplary scenario of Figure 3, two separate source nodes o1 and o2 are implemented, while one destination node functions as either destination node d1 or destination node d2. In this way, two traffic requests are defined, one being the traffic request between source o1 and destination d1, and the other being the traffic request between another source o2 and the same destination d2. The determined data volume is transferred between o1, d1 and o2, d2. Here, the central optimization problem is to select and determine the optimal communication paths for traffic requests o1, d1 and o2, d2 such that the shortest possible communication paths are selected for the shortest path approach as much as possible, and at the same time, the overall utilization rate of the capacity within the network is minimized uniformly within the network, i.e., the overall capacity utilization rate of the connections within the possible communication paths is minimized.

[0063] According to Figure 3, a set of possible short communication paths p1 - p4 is specified in advance. This can be done, for example, through the application of Dijkstra's algorithm that calculates the possible short communication paths between two pairs of source nodes o1, o2 and destination nodes d1, d2 for each traffic request to be transferred. As exemplarily shown in Figure 3, path p1 goes from o1 to d1 via segment node s2. Path p2 goes from o1 to d1 via segment node s1. Path p3 goes from o2 to d2 via segment node s2. Path p4 goes from o2 to d2 via segment node s1. These are possible communication paths for routing data traffic from their respective sources o1 and o2 towards destinations d1 / d2.

[0064] Segment nodes s1 and s2 function as relay nodes in their respective communication paths. Despite the central optimization problem of calculating short communication paths optimized for each traffic demand, a special aspect of this optimization problem is to select segment node s for the transmission of traffic demands such that the central optimization problem is still satisfied. The main advantage of the configuration of segment nodes s1 and s2 is that it provides freedom and flexibility in routing traffic demands. As described above, FIG. 3 exemplarily implements the two-segment routing (2SR) protocol.

[0065] FIG. 3 is further referred to as edges e1 and e2 and further shows two exemplary connections that may be within a possible communication path. Edge e1 is configured between segment node s1 and destinations d1 / d2, and edge e2 is configured between segment node s2 and destinations d1 / d2.

[0066] For the routing of traffic requests between the starting points o1, o2 and the destinations d1 / d2, different options are assumed. The routing option from o1 to d1 is the path p1 such that the data traffic o1, d1 is transferred via the edge e2. Another routing option from o1 to d1 is the path p2 such that the data traffic o1, d1 is transferred via the edge e1. Similar assumptions can apply to the data traffic o2, d2. Here, the first option is the path p3 such that the data traffic o2, d2 is routed via the edge e2. The second option for routing from o2 to d2 is the path p4 such that the data traffic o2, d2 is routed via the edge e1. As can be seen from these different options for the routing of the data traffic o1, d1 and o2, d2, there are combinations of communication paths for o1, d1 and o2, d2, and each of the two edges e1 and e2 bears the burden of only one traffic request. This is given, for example, by o1, d1 passing through path p1 and o2, d2 passing through path p4. However, there are also possible combinations of communication paths where one of the edges e1 and e2 bears the substantial and heavy load of both traffic requests and the other of the edges e1 and e2 is not used at all. This is given, for example, by o1, d1 passing through path p1 and o2, d2 passing through path p3 (or o1, d1 passing through path p2 and o2, d1 passing through path p4).

[0067] The latter combination has the significant drawback that the capacity utilization rate of one of the edges e1 and e2 is extremely high, which may lead to overload or failure of each edge. Therefore, the optimization problem lies in determining and selecting communication paths for the traffic requests o1, d1 and o2, d2 such that the overall capacity utilization rate is distributed across both of the edges e1 and e2. In a particular implementation of such an optimization problem, the respective assignment of the segment nodes s1 and s2 is performed for the optimized selection of the respective communication paths for the transfer of o1, d1 and o2, d2.

[0068] To solve such an optimization problem, the partial capacity utilization rate of all edges within a set of potentially short communication paths can be calculated for the entire set of traffic demands. As exemplarily given in FIG. 3, such a strategy involves calculating the partial capacity utilization rate of each of edges e1 and e2 for each of traffic demands o1, d1 and o2, d2. The “partial capacity utilization rate” of each edge means that, based on the respective utilization capacity limit of each edge, the portion of the capacity utilization rate required for each traffic demand transmitted through this edge is calculated.

[0069] For example, with respect to FIG. 3, assume that each traffic demand o1, d1 and o2, d2 requires half of the maximum utilization capacity of each of edges e1 and e2 (i.e., 50% of the capacity). This means that each of edges e1 and e2 is burdened with half of the utilized capacity for each of traffic demands o1, d1 and o2, d2. In other words, for example, if o1, d1 passes through path p1 and o2, d2 passes through path p4, both edges e1 and e2 are burdened with 50% of their utilization capacity limits. Otherwise, for example, if o1, d1 passes through path p1 and o2, d2 passes through path p3, edge e2 is fully and completely burdened (2 x 50% = 100%), thereby reaching its capacity limit and resulting in the use of the entire edge capacity of e2. A similar assumption can be applied to edge e1 when o1, d1 passes through path p2 and o2, d2 passes through p4.

[0070] Such a calculation of the partial capacity utilization rate is performed for all the remaining edges within potentially short communication paths p1 to p4 in the scenario of FIG. 3. The calculated partial capacity utilization rate is then formulated as a term of a secondary stress function, as detailed below and with reference to FIG. 4C.

[0071] FIGS. 4A to 4D show an exemplary mathematical formulation of a partial optimization problem according to the approach as described above with respect to FIGS. 2B and 3. The mathematical formulation in FIGS. 4A to 4D is represented as a so-called Hamiltonian function, a short Hamiltonian.

[0072] The mathematical formulation of FIG. 4A formulates the routing condition that each traffic demand between each origin and destination (o, d) is routed along the shortest path or through exactly one segment node between each origin node o and each destination node d. For the exemplary scenario of FIG. 3, this means that traffic o1, d1 and traffic o2, d2 can be routed along the shortest path (not explicitly shown in FIG. 3) or through exactly one segment node s1, s2. This means that for each of the traffic demands o1, d1 and o2, d2, only one segment node s1 or s2 can be selected for each demand.

[0073] The mathematical formulation of FIG. 4a can assume a value of "0" or a value of "1" (or both with a specific probability), and is a binary variable represented as a bit (or a q-bit as used below) in a quantum concept processor.

Number

Number

Number

Number

Number

[0074] The mathematical formulation in Figure 4B formulates a cost condition regarding the total number of segment nodes used within network 1. This Hamiltonian sums all Q - bits having the value "1" (indicating the segment nodes selected for routing) over all traffic demands o, d

Number

[0075] The mathematical formulation of the Hamiltonian according to Figure 4C represents the core optimization problem formulated as a quadratic stress function considering the calculated partial capacity utilization rates of all edges e within the set p of possible short communication paths for all traffic demands o, d. Thus, the core optimization problem here is to minimize the Hamiltonian according to Figure 4C in order to find communication paths optimized for all traffic demands within the network.

[0076] The Hamiltonian of FIG. 4C takes into account all traffic demands o, d, and further takes into account the set s of segment nodes that can be selected to select a communication path p, and considers the sum terms for each edge e in the possible communication path p. Assuming that the equation of FIG. 4A is satisfied, the Hamiltonian of FIG. 4C describes the sum of all calculated partial capacity utilization rates of all edges e that are part of the possible communication path p

Number

Number

Number

Number

Number

Number

Number

[0077] The Hamiltonian of FIG. 4C has two main terms t1 and t2. t1 formulates routing by the 2SR approach, and the first equation of t1 is the partial capacity utilization rate of all demands o, d in the segment between the origin o and the segment node s [Number] Sum them up. Next, the second equation of t1 is all the requests o in the segment between each segment node s and the destination d, and the fractional capacity utilization rate of all d [Number] Sum them up. The second term t2 is the partial capacity utilization rate of all edges e within these shortest paths [Number] By summing them up, consider the shortest paths for all requests o, d. The shortest path is defined such that possible segment nodes s are not selected.

[0078] The Hamiltonian of Figure 4C is formulated such that either term t1 or term t2 (but not both) is considered. Therefore, it is possible to select a path with one segment node s or the shortest path without segment node s. This is due to the setting of each qubit [Number] If at least one segment node s is selected, at least one qubit [Number] has the value "1". In this case, term t1 multiplied by the qubit [Number] (one of which has the value "1") is considered. However, in this case, since term t21 of t2 becomes "0", the second term t2 becomes "0". On the contrary, when the segment node s is not selected, all qubits [Number] has the value "0", and the term t1 is multiplied by "0", resulting in being not considered. However, in this case, the second term t2 is considered because the internal expression t21 becomes "1".

[0079] The exponent parameter "q" which should have a value greater than 0 and less than or equal to 1 can be used as an additional problem-specific control parameter. When q < 1, smaller volume requirements are preferentially assigned to segment nodes that result in a set of the shortest paths with small edge capacities. In contrast, when q = 1, all requirements are evenly distributed without further priority.

[0080] In this way, the Hamiltonian of FIG. 4C is formulated to consider either the shortest path or exclusively one segment node s for each request o, d to select the communication path p, thereby considering the partial capacity utilization rate of all edges e in each respective part of the possible communication path p

Number

[0081] Considering the scenario of FIG. 3 for an example of two edges e1 and e2, the Hamiltonian according to FIG. 4C can have the following expression when q = 1.

Number

[0082] Under the assumption that each request o1, d1 and o2, d2 places a burden of half (50%) of its capacity on the respective edges e1, e2, as described above, when o1, d1 and o2, d2 are routed through different segment nodes s, the above terms reach the minimum value. And the above terms are as follows.

Number

[0083] Otherwise, if o1, d1 and o2, d2 are routed through one common segment node s1 or s2 (no other segment nodes are used), the above terms are as follows.

Number

[0084] The above example shows that selecting different segment nodes s for two requests o1, s1 and o2, d2 is a preferred solution for distributing the overall capacity utilization rate across the network and achieving the optimization goal of short-path routing.

[0085] The Hamiltonian of Figure 4C is generally solved by a quantum theory concept processor that executes through different settings of the values of each qubit

Number

Number

Number

Number

[0086] FIG. 4D finally shows the global QUBO formulation of the overall optimization problem, in which the three partial optimization problems according to FIGS. 4A-4C are each multiplied by their respective weighting factors A, B, C and summed up towards the global optimization problem. This global optimization problem is finally processed by applying a computer-implemented algorithm within a quantum concept processor. In this regard, the minimization of the Hamiltonian according to FIG. 4C is performed, whereby further optimization constraints formulated in the Hamiltonian according to FIGS. 4A and 4B are taken into account.

[0087] In particular, with respect to the Hamiltonian of FIG. 4B, a compromise is calculated between the minimum value of the Hamiltonian of FIG. 4C and the minimum value of the cost depending on the number of segment nodes s used, formulated in the Hamiltonian of FIG. 4B. The global minimum value of the Hamiltonian of FIG. 4C can be found by significantly increasing the number of segment nodes s selected and used. However, this significantly increases the deployment and maintenance costs of the segment nodes s, as represented by the Hamiltonian of FIG. 4B. In contrast, strictly minimizing the cost for the deployment and maintenance of the segment nodes s (in fact, not selecting any segment nodes at all) would lead to the fact that the Hamiltonian of FIG. 4C cannot be solved satisfactorily to reach a sufficient minimum value.

[0088] The global optimization problem formulated in FIG. 4D takes into account minimizing the total number of segment nodes used for routing (FIG. 4B), and further considering the constraint of selecting only one segment node for each traffic demand or not being able to select any at all (FIG. 4A), and tries to find the minimum value of the quadratic stress function with respect to the capacity utilization rate (FIG. 4C), respecting the determined compromise points among these partial optimization problems.

[0089] With the help of the weighting coefficients A, B, and C according to FIG. 4D, different weightings and foci on different partial optimization problems can be set. For example, when coefficient B is larger than coefficient C, the focus is on minimizing the maximum capacity utilization rate and the uniform distribution of the capacity utilization rate. Otherwise, when coefficient B is smaller than coefficient C, the focus is more on minimizing the overall cost of the segment nodes used. In addition, since coefficient A can be set very high, for example, this constraint is not substantially violated during optimization. As an alternative to the mathematical formulation A > B, C > 0, for example, another mathematical formulation is A > B > 0, C >= 0.

[0090] FIG. 5 shows an exemplary schematic diagram of an algorithm that executes the approach described above. FIG. 5 shows the processing of the steps and procedures of the method described above considering a set 5 of traffic demands, and these traffic demands 5 are processed in a decomposed manner. This has advantages or is even necessary for processing the described method despite the fact that the hardware performance of the quantum concept processor 6 for solving the optimization problem is limited as described above with respect to FIGS. 4A to 4D. The quadratic optimization problem is very complex and has to be decomposed into several partial solutions that can be processed iteratively to find the optimal solution. In each iteration, one partial solution is found by the quantum concept processor 6.

[0091] Therefore, to decompose the problem, traffic requirement 5 is sorted in descending order by volume and split into separate parts of the requirement. In particular, traffic requirements 5 with higher volumes can be considered first, followed by requirements with smaller volumes. For this purpose, in the implementation of the algorithm according to FIG. 5, first, a subset 7 of traffic requirement 5 is considered. For example, subset 7 of traffic requirement 5 includes a specific percentage of the total volume among the entire set of traffic requirements 5. For example, subset 7 includes 30% of the traffic volume transmitted over the network. In the next step, for this subset 7 of traffic requirement 5, a subset 8 of possible segment nodes s is selected from the set of possible segment nodes s. In this way, the total number of possible segment nodes s is reduced to a subset 8 of the most relevant possible segment nodes s with respect to the selected subset 7 of traffic requirement 5. For example, the subset 8 of possible segment nodes s for each traffic requirement 5 can be pre-selected as the segment nodes s that are very close to the shortest path between the respective source node and destination node of each traffic requirement 5 that is part of subset 7.

[0092] The pre-processed subset 7 of traffic requirement 5 with the selected subset 8 of segment nodes s is then input into the algorithm procedure within the quantum concept processor 6. For example, the quantum concept processor 6 according to FIG. 5 is configured to solve the optimization problem by quantum annealing emulation. The quantum concept processor 6 applies the mathematical formulation of the overall optimization problem according to FIG. 4D. The quantum concept processor 6 then uses the selected subset 8 of segment nodes s to calculate, according to FIG. 4D, the optimized solution of the global optimization problem for the subset 7 of traffic requirement 5.

[0093] After the algorithm procedure is completed, the finally calculated minimum value of the global optimization problem according to FIG. 4D is output from the quantum concept processor 6 for each subset 7 of the traffic demands 5. Next, the communication path p determined by the found optimal value of the optimization problem is stored for the subset 7 of the traffic demands 5.

[0094] Furthermore, considering that the determined and optimized routing for the subset 7 of the traffic demands 5 already requires a certain amount of capacity within the network, the remaining usage capacity limits of each edge within the set of possible short communication paths p are updated. If any requests remain, the procedure is repeatedly executed for the remaining traffic demands 5 until all traffic demands 5 of the set of traffic demands are processed. In this case, the maximum capacity utilization rate of all traffic demands 5 transferred via the communication path determined through the network is calculated. Then, the algorithm ends.

[0095] Thus, by applying the computer-implemented algorithm procedure according to FIG. 5, based on the above implementation and description regarding FIGS. 2B - 4D, optimized routing can be provided for all traffic demands on the short communication paths individually selected via the communication network 1.

[0096] The formulation of the optimization problem as a QUBO representation has elegant properties with respect to the quantum concept computing applied herein within the processor 6. Today, quantum concept computing still reaches great limitations. However, as computer science increasingly develops towards quantum computers, the approach described herein may be further enhanced and developed in the future. For example, when quantum computing becomes increasingly applicable to the increasing complexity of the underlying optimization problem, the decomposition strategy explained from the perspective of FIG. 5 can be increasingly reduced, which means that the optimization problem can be increasingly processed and calculated as a whole without decomposition steps and iterations of the problem. Further, as quantum computing becomes increasingly applicable, an increasing number of qubits, increasingly complex optimization problems, and / or increasingly non-linear constraints can be considered by the approach described herein.

[0097] The approach described herein is mainly applicable to communication networks. However, this approach can also be applied to any other network, such as a railway network, an energy grid, a transportation network, etc., where a specific "traffic" or "load" has to be transmitted across the network via an optimized route.

[0098] The embodiments shown and described herein are merely illustrative.

Description of Signs

[0099] 1 Communication network 2, 2a~2f Communication nodes 3, 3a~3e Aggregation nodes 4 Connection between adjacent nodes 5 Set of traffic requirements 5a, 5b Traffic requirements 6 Quantum concept processor 7 Subset of traffic requirements 8 Subset of segment nodes d, d1, d2 Destination nodes e, e1, e2 edges o, o1, o2 starting nodes p, p1~p4 possible communication paths s set of segment nodes s1, s2 segment nodes t1, t2, t21 terms of the secondary stress function

Claims

1. In a communication network (1) having a plurality of communication nodes (2) connectable via an edge (4) of a communication path for routing data traffic, a computer-implemented method for optimizing the routing of the data traffic, comprising: Capturing a set (5) of traffic requests, each traffic request (5a, 5b) specifying the transfer of a determined data volume from a source node (o) to a destination node (d) among the plurality of communication nodes (2); Specifying a set (p) of communication paths that may be short among the possible communication paths between each respective source node (o) and each respective destination node (d) specified in the set (5) of traffic requests, and assigning a respective utilization capacity limit to the edge (e) within the set (p) of communication paths that may be short; Calculating, for the set (5) of traffic requests, a partial capacity utilization rate of the edge (e) within the set (p) of communication paths that may be short, the partial capacity utilization rate being calculated based on the respective utilization capacity limits; Formulating the calculated partial capacity utilization rate as a term of a quadratic stress function; Using a quantum theory concept processor (6) to select, for each traffic request (5a, 5b) of the set (5) of traffic requests, one short communication path (p1-p4) from the set (p) of communication paths that may be short such that the quadratic stress function is minimized, thereby determining an optimized routing.

2. Specifying a set (s) of possible segment nodes among the plurality of communication nodes (2), each of the possible segment nodes (s1, s2) defining a communication path that may be short (p1-p4) as an intermediate node as a member of the set (p) of communication paths that may be short between the source node (o) and the destination node (d); In the secondary stress function, formulating segment node terms, where each segment node term connects the calculated partial capacitance utilization rate of the edge (e) of each short possible communication path (p1 - p4) to those segment nodes (s1, s2) within the set (s) of possible segment nodes connected to each short possible communication path (p1 - p4). Using the quantum theory concept processor (6), calculating the segment node terms so that the secondary stress function is minimized for determining the optimized routing, and selecting segment nodes (s1, s2) within the set (s) of possible segment nodes, the method according to claim 1 further comprising this.

3. The segment node terms are calculated considering the path condition that each traffic requirement in the set (5) of traffic requirements is routed along the shortest path or through exactly one segment node (s1, s2) between each source node and each destination node, the method according to claim 2.

4. The segment node terms are calculated considering the path condition that each traffic requirement (5a, 5b) in the set (5) of traffic requirements is routed along the shortest path or through a plurality of segment nodes (s1, s2) between each source node (o) and each destination node (d), the method according to claim 2.

5. The segment node terms are calculated considering cost conditions so that the number of selected segment nodes (s1, s2) is minimized, the method according to claim 2.

6. The secondary stress function is formulated as a quadratic unconstrained binary optimization (QUBO) function, the method according to claim 1.

7. The segment node terms are calculated considering cost conditions so that the number of selected segment nodes (s1, s2) is minimized, The method according to claim 3, wherein at least one of the secondary stress function, the path condition, and the cost condition is weighted and coupled to a global QUBO function, respectively.

8. selecting a subset (7) of traffic requests from the set (5) of traffic requests; executing a method on the subset (7) of traffic requests; storing the determined and optimized routing for the subset (7) of traffic requests; updating the remaining utilization capacity limit of each of the edges (e) in the set of possible short communication paths, taking into account the determined and optimized routing for the subset (7) of traffic requests, the method according to claim 1 further comprising.

9. selecting a subset (8) of possible segment nodes from the set (s) of possible segment nodes; the method according to claim 2, further comprising executing the method on the subset (8) of possible segment nodes.

10. The method according to claim 8, wherein the method is repeatedly executed on the remaining traffic requests (5a, 5b) until all traffic requests (5a, 5b) in the set (5) of traffic requests are processed.

11. A quantum concept processor (6), particularly a digital annealing processing unit or a quantum annealing processing unit, configured to execute one or more steps of the method according to claim 1.

12. A computer program comprising instructions which, when the computer program is executed by one or more processors, cause each of the one or more processors to execute one or more steps of the method according to claim 1.

13. A computer-readable storage medium storing the computer program according to claim 12. **Claim 14** A workplace for a network planner configured to verify an optimized routing determined by the method according to claim 1. **Claim 15** An interface configuration including one or more interfaces to a plurality of communication nodes (2) of a communication network (1) through which data traffic is routed, the interface configuration being configured to automatically deploy an optimized routing determined by the method according to claim 1 to the communication nodes (2) of the communication network (1).

Images