Machine learning segment routing for multiple traffic matrices
In some embodiments, there may be provided a method that includes receiving a first traffic matrix; receiving information regarding links associated with each segment of the network; determining a total amount of segment flow using the at least one non-linear deflection parameter applied to the traffic demand of the first traffic matrix; determining a link flow for each of the links using the total amount of segment flow and the second input to the machine learning model; determining link utilization for each of the links using the link flows and a capacity for each of the links; learning, by the machine learning model using a gradient descent, a minimum of a maximum amount of the link utilization over the links by at least adjusting a value of the at least one non-linear deflection parameter. Related systems, methods, and articles of manufacture are also disclosed.
1. A method comprising:
receiving, as a first input to a machine learning model, a first traffic matrix, wherein the first traffic matrix indicates traffic demand between at least a source node and a destination node of a network;
receiving, as a second input to the machine learning model, information regarding links associated with each segment of the network, wherein each segment corresponds to a pair of nodes of the network;
transforming, by the machine learning model, at least one deflection parameter into at least one non-linear deflection parameter, wherein the at least one deflection parameter indicates a fractional amount of traffic that is carried between the source node and the destination node and deflected to an intermediary node;
determining, by the machine learning model, a total amount of segment flow using the at least one non-linear deflection parameter applied to the traffic demand of the first traffic matrix;
determining, by the machine learning model, a link flow for each of the links using the total amount of segment flow and the second input to the machine learning model;
determining, by the machine learning model, link utilization for each of the links using the link flows and a capacity for each of the links;
learning, by the machine learning model using a gradient descent, a minimum of a maximum amount of the link utilization over the links by at least adjusting a value of the at least one non-linear deflection parameter; and
providing, by the machine learning model, an output comprising the at least one non-linear deflection parameter and/or the at least one non-linear deflection parameter transformed back into a linear domain.
2. The method of claim 1 , further comprising:
receiving a plurality of traffic matrices indicative of traffic demands between at least the source node and the destination node of the network.
3. The method of claim 2 , wherein the total amount of segment flow is determined using the at least one non-linear deflection parameter applied to the traffic demands obtained from the plurality of traffic matrices.
4. The method of claim 1 , wherein the output comprising the at least one non-linear deflection parameter and/or the at least one non-linear deflection parameter transformed back into the linear domain takes into account the learning using a plurality of traffic matrices indicative of traffic demands between at least the source node and the destination node of the network.
5. The method of claim 1 , wherein the output is provided to one or more routing tables of the network.
6. The method of claim 1 , wherein the output provides the fractional amount of traffic that is deflected to the intermediary node.
7. The method of claim 1 , wherein the at least one non-linear deflection parameter is generated by at least transforming the at least one deflection parameter into an exponential domain and then normalizing the at least one non-linear deflection parameter to a value between zero and one.
8. The method of claim 1 , wherein the machine learning model comprises a loss function including the at least one non-linear deflection parameter.
9. The method of claim 1 , wherein the machine learning model performs the learning as an unconstrained optimization of the minimum of the maximum amount of the link utilization.
10. An apparatus comprising:
at least one processor;
at least one memory including instructions, which when executed by the at least one processor causes the apparatus to cause operations comprising:
receiving, as a first input to a machine learning model, a first traffic matrix, wherein the first traffic matrix indicates traffic demand between at least a source node and a destination node of a network;
receiving, as a second input to the machine learning model, information regarding links associated with each segment of the network, wherein each segment corresponds to a pair of nodes of the network;
transforming, by the machine learning model, at least one deflection parameter into at least one non-linear deflection parameter, wherein the at least one deflection parameter indicates a fractional amount of traffic that is carried between the source node and the destination node and deflected to an intermediary node;
determining, by the machine learning model, a total amount of segment flow using the at least one non-linear deflection parameter applied to the traffic demand of the first traffic matrix;
determining, by the machine learning model, a link flow for each of the links using the total amount of segment flow and the second input to the machine learning model;
determining, by the machine learning model, link utilization for each of the links using the link flows and a capacity for each of the links;
learning, by the machine learning model using a gradient descent, a minimum of a maximum amount of the link utilization over the links by at least adjusting a value of the at least one non-linear deflection parameter; and
providing, by the machine learning model, an output comprising the at least one non-linear deflection parameter and/or the at least one non-linear deflection parameter transformed back into a linear domain.
11. The apparatus of claim 10 , further comprising:
receiving a plurality of traffic matrices indicative of traffic demands between at least the source node and the destination node of the network.
12. The apparatus of claim 11 , wherein the total amount of segment flow is determined using the at least one non-linear deflection parameter applied to the traffic demands obtained from the plurality of traffic matrices.
13. The apparatus of claim 10 , wherein the output comprising the at least one non-linear deflection parameter and/or the at least one non-linear deflection parameter transformed back into the linear domain takes into account the learning using a plurality of traffic matrices indicative of traffic demands between at least the source node and the destination node of the network.
14. The apparatus of claim 10 , wherein the output is provided to one or more routing tables of the network.
15. The apparatus of claim 10 , wherein the output provides the fractional amount of traffic that is deflected to the intermediary node.
16. The apparatus of claim 10 , wherein the at least one non-linear deflection parameter is generated by at least transforming the at least one deflection parameter into an exponential domain and then normalizing the at least one non-linear deflection parameter to a value between zero and one.
17. The apparatus of claim 10 , wherein the machine learning model comprises a loss function including the at least one non-linear deflection parameter.
18. The apparatus of claim 10 , wherein the machine learning model performs the learning as an unconstrained optimization of the minimum of the maximum amount of the link utilization.
19. A non-transitory computer-readable storage medium including instructions, which when executed by at least one processor causes operations comprising:
receiving, as a first input to a machine learning model, a first traffic matrix, wherein the first traffic matrix indicates traffic demand between at least a source node and a destination node of a network;
receiving, as a second input to the machine learning model, information regarding links associated with each segment of the network, wherein each segment corresponds to a pair of nodes of the network;
transforming, by the machine learning model, at least one deflection parameter into at least one non-linear deflection parameter, wherein the at least one deflection parameter indicates a fractional amount of traffic that is carried between the source node and the destination node and deflected to an intermediary node;
determining, by the machine learning model, a total amount of segment flow using the at least one non-linear deflection parameter applied to the traffic demand of the first traffic matrix;
determining, by the machine learning model, a link flow for each of the links using the total amount of segment flow and the second input to the machine learning model;
determining, by the machine learning model, link utilization for each of the links using the link flows and a capacity for each of the links;
learning, by the machine learning model using a gradient descent, a minimum of a maximum amount of the link utilization over the links by at least adjusting a value of the at least one non-linear deflection parameter; and
providing, by the machine learning model, an output comprising the at least one non-linear deflection parameter and/or the at least one non-linear deflection parameter transformed back into a linear domain.