Diversifying k-shortest path computation results for multiple vertices as destination or as inclusion constraint
Systems and methods for diversifying k-shortest path computation results for multiple vertices as destination or as inclusion constraint include receiving a request for k paths, in a network from a source port at a source node to a node, wherein the node includes a plurality of ports; representing the network as a graph with vertices representing ports, edges representing connections between the ports, and with weights assigned to each of the edges; assigning a dummy destination vertex to connect to each of the plurality of ports and associated edges between each of the plurality of ports and the dummy destination vertex having a same weight; determining up to the k paths between the source port and the dummy destination vertex, and, subsequent to the determining each of the k paths, assigning a corresponding edge of the associated edges in each of the k paths a predetermined value.
1 . A non-transitory computer-readable medium comprising instructions that, when executed, cause one or more processors to implement steps of:
receiving a request for k paths, k>1, in a network from a source port at a source node to a node, wherein the node includes a plurality of ports;
representing the network as a graph with vertices representing ports including the source port and the plurality of ports, edges representing connections between the ports, and with weights assigned to each of the edges;
assigning a dummy destination vertex to connect to each of the plurality of ports and associated edges between each of the plurality of ports and the dummy destination vertex having a same weight; and
determining up to the k paths between the source port and the dummy destination vertex, and, subsequent to the determining each of the k paths, assigning a corresponding edge of the associated edges in each of the k paths a predetermined value, greater than the same weight, wherein the predetermined value is set such that subsequent paths are biased toward different ones of the plurality of ports to ensure diversity across the plurality of ports at the node.
2 . The non-transitory computer-readable medium of claim 1 , wherein the steps further include
removing the dummy destination vertex and the associated edges from the k paths to form k shortest paths having diversity across the plurality of ports at a destination node; and
providing the k shortest paths in response to the request.
3 . The non-transitory computer-readable medium of claim 1 , wherein the node is an intermediate node that is an inclusion constraint and the request further includes a destination port at a destination node, and wherein the steps include
determining k paths from associated ports of the plurality of ports at the intermediate node to the destination port; and
returning k shortest paths based on the k paths between the source port and the intermediate node and the k paths from associated ports of the plurality of ports at the intermediate node to the destination port.
4 . The non-transitory computer-readable medium of claim 1 , wherein the node is a destination node, such that the k paths include diversity across the plurality of ports at the destination node.
5 . The non-transitory computer-readable medium of claim 1 , wherein the predetermined value is at least an order of magnitude greater than the weights.
6 . The non-transitory computer-readable medium of claim 1 , wherein the determining utilizes Yen's algorithm.
7 . The non-transitory computer-readable medium of claim 1 , wherein the network is an optical network, and the ports represent wavelength connections.
8 . The non-transitory computer-readable medium of claim 1 , wherein the network operates at one or more of Layers 1, 2, and 3, and the ports represent either Time Division Multiplexing (TDM) or packet connections.
9 . The non-transitory computer-readable medium of claim 1 , wherein the one or more processors are in a Path Computation Engine (PCE).
10 . A method comprising steps of:
receiving a request for k paths, k>1, in a network from a source port at a source node to a node, wherein the node includes a plurality of ports;
representing the network as a graph with vertices representing ports including the source port and the plurality of ports, edges representing connections between the ports, and with weights assigned to each of the edges;
assigning a dummy destination vertex to connect to each of the plurality of ports and associated edges between each of the plurality of ports and the dummy destination vertex having a same weight; and
determining up to the k paths between the source port and the dummy destination vertex, and, subsequent to the determining each of the k paths, assigning a corresponding edge of the associated edges in each of the k paths a predetermined value, greater than the same weight, wherein the predetermined value is set such that subsequent paths are biased toward different ones of the plurality of ports to ensure diversity across the plurality of ports at the node.
11 . The method of claim 10 , wherein the steps further include
removing the dummy destination vertex and the associated edges from the k paths to form k shortest paths having diversity across the plurality of ports at a destination node; and
providing the k shortest paths in response to the request.
12 . The method of claim 10 , wherein the node is an intermediate node that is an inclusion constraint and the request further includes a destination port at a destination node, and wherein the steps include
determining k paths from associated ports of the plurality of ports at the intermediate node to the destination port; and
returning k shortest paths based on the k paths between the source port and the intermediate node and the k paths from associated ports of the plurality of ports at the intermediate node to the destination port.
13 . The method of claim 10 , wherein the node is a destination node, such that the k paths include diversity across the plurality of ports at the destination node.
14 . The method of claim 10 , wherein the predetermined value is at least an order of magnitude greater than the weights.
15 . The method of claim 10 , wherein the determining utilizes Yen's algorithm.
16 . The method of claim 10 , wherein the network is an optical network, and the ports represent wavelength connections.
17 . The method of claim 10 , wherein the network operates at one or more of Layers 1, 2, and 3, and the ports represent either Time Division Multiplexing (TDM) or packet connections.
18 . The method of claim 10 , wherein the method is implemented in a Path Computation Engine (PCE).
19 . A system comprising:
one or more processors; and
memory storing instructions that, when executed, cause the one or more processors to
receive a request for k paths, k>1, in a network from a source port at a source node to a node, wherein the node includes a plurality of ports,
represent the network as a graph with vertices representing ports including the source port and the plurality of ports, edges representing connections between the ports, and with weights assigned to each of the edges,
assign a dummy destination vertex to connect to each of the plurality of ports and associated edges between each of the plurality of ports and the dummy destination vertex having a same weight, and
determine up to the k paths between the source port and the dummy destination vertex, and, subsequent to the each of the k paths being determined, assigning a corresponding edge of the associated edges in each of the k paths a predetermined value, greater than the same weight, wherein the predetermined value is set such that subsequent paths are biased toward different ones of the plurality of ports to ensure diversity across the plurality of ports at the node.
20 . The system of claim 19 , wherein the instructions that, when executed, further cause the one or more processors to
remove the dummy destination vertex and the associated edges from the k paths to form k shortest paths having diversity across the plurality of ports at a destination node, and
provide the k shortest paths in response to the request.