Selective disaggregation advertisement for routing in multi-plane topologies
Techniques for identifying nodes in a data center fabric that are affected by a failure in the fabric, and selectively sending disaggregation advertisements to the nodes affected by the failure. The techniques include a process where a component monitors the network fabric to identify communication paths between leaf nodes, and determines what leaf nodes would be affected by a failure in those communication paths. The component may detect a failure in the network and determine which communication paths, and thus which leaf nodes, are affected by the failure and send disaggregation advertisements to the affected leaf nodes. In some examples, ingress leaf nodes send data through the fabric that indicate egress nodes for the communication paths. Intermediate nodes along may receive the data from the leaf nodes to identify communication paths, and the notify only affected nodes upon detecting a failure in the network.
1. A method for an orchestration component to manage data flows in a multi-plane network, the method comprising:
identifying, by the orchestration component, a failure in a first plane of the multi-plane network;
determining, by the orchestration component, that communication paths to a first leaf node are affected by the failure in the first plane;
identifying a first set of nodes in the multi-plane network that have first communication paths that are affected by the failure, wherein a second set of nodes in the multi-plane network have second communication paths that are unaffected by the failure;
sending disaggregation advertisements indicative of the failure to the first set of nodes while refraining from sending the disaggregation advertisements to the second set of nodes;
identifying a second leaf node having a first communication path to the first leaf node that is affected by the failure in the first plane; and
causing the second leaf node to communicate with the first leaf node using a second communication path in a second plane of the multi-plane network.
2. The method of claim 1 , further comprising:
receiving, at the orchestration component, telemetry data from a plurality of devices in the multi-plane network;
analyzing the telemetry data to determine a plurality of communication paths between a plurality of pairs of leaf nodes in the multi-plane network; and
generating mappings between individual pairs of leaf nodes of the plurality of pairs of leaf nodes based on the individual pairs of leaf nodes having at least one communication path,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying, from the mappings, that the second leaf node is mapped to the first leaf node.
3. The method of claim 1 , further comprising:
receiving, from the second leaf node, a request to monitor the first communication path to the first leaf node; and
storing a mapping between the second leaf node and the first leaf node based at least in part on receiving the request,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying the mapping.
4. The method of claim 1 , further comprising:
receiving a request to establish a secure tunnel between the first leaf node and the second leaf node;
causing the first leaf node and the second leaf node to establish the secure tunnel over the first communication path;
storing a mapping between the second leaf node and the first leaf node based at least in part on the secure tunnel being established between the second leaf node and the first leaf node,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying the mapping.
5. The method of claim 1 , further comprising:
receiving, at the orchestration component, telemetry data representing communications between a plurality of devices in the multi-plane network over a period of time;
analyzing the telemetry data to determine that a third leaf node has previously established a third communication path with the first leaf node in the first plane during the period of time, wherein the third leaf node is not currently communicating with the first leaf node;
sending an instruction to the third leaf node indicating that the third leaf node either (i) cannot reach the first leaf node via the first plane, or (ii) that the third leaf node can communicate with the first leaf node via the second plane.
6. The method of claim 1 , wherein causing the second leaf node to communicate using the second communication path to the first leaf node using the second plane includes:
sending a control-plane instruction to the second leaf node that instructs the second leaf node to refrain from load balancing communication flows to the first leaf node through the first plane.
7. The method of claim 1 , further comprising:
identifying a third leaf node having a third communication path to the first leaf node that is affected by the failure in the first plane;
determining that the second leaf node has a first number of communication flows with the first leaf node;
determining that the third leaf node has a second number of communication flows with the first leaf node, wherein the first number of communication flows is greater than the second number of communication flows; and
notifying the second leaf node of the failure in the first plane prior to notifying the third leaf node based at least in part on the first number of communication flows being greater than the second number of communication flows.
8. The method of claim 1 , further comprising:
identifying a third leaf node having a third communication path to the first leaf node that is affected by the failure in the first plane;
determining that the second leaf node has a first communication flow with the first leaf node associated with a first quality of service (QoS) metric;
determining that the third leaf node has a second communication flow with the first leaf node associated with a second QoS metric; and
notifying the second leaf node of the failure in the first plane prior to notifying the third leaf node based at least in part on the first QoS metric and the second QoS metric.
9. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
identifying, by an orchestration component, a failure in a first plane of a multi-plane network;
determining, by the orchestration component, that communication paths to a first leaf node are affected by the failure in the first plane;
identifying a first set of nodes in the multi-plane network that have first communication paths that are affected by the failure, wherein a second set of nodes in the multi-plane network have second communication paths that are unaffected by the failure;
notifying the first set of nodes indicating the failure while refraining from notifying the second set of nodes about the failure;
identifying a second leaf node having a first communication path to the first leaf node that is affected by the failure in the first plane; and
causing the second leaf node to communicate with the first leaf node using a second communication path in a second plane of the multi-plane network.
10. The system of claim 9 , the operations further comprising:
receiving, at the orchestration component, telemetry data from a plurality of devices in the multi-plane network;
analyzing the telemetry data to determine a plurality of communication paths between a plurality of pairs of leaf nodes in the multi-plane network; and
generating mappings between individual pairs of leaf nodes of the plurality of pairs of leaf nodes based on the individual pairs of leaf nodes having at least one communication path,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying, from the mappings, that the second leaf node is mapped to the first leaf node.
11. The system of claim 9 , the operations further comprising:
receiving, from the second leaf node, a request to monitor the first communication path to the first leaf node; and
storing a mapping between the second leaf node and the first leaf node based at least in part on receiving the request,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying the mapping.
12. The system of claim 9 , the operations further comprising:
receiving a request to establish a secure tunnel between the first leaf node and the second leaf node;
causing the first leaf node and the second leaf node to establish the secure tunnel over the first communication path;
storing a mapping between the second leaf node and the first leaf node based at least in part on the secure tunnel being established between the second leaf node and the first leaf node,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying the mapping.
13. The system of claim 9 , the operations further comprising:
receiving, at the orchestration component, telemetry data representing communications between a plurality of devices in the multi-plane network over a period of time;
analyzing the telemetry data to determine that a third leaf node has previously established a third communication path with the first leaf node in the first plane during the period of time, wherein the third leaf node is not currently communicating with the first leaf node;
sending an instruction to the third leaf node indicating that the third leaf node either (i) cannot reach the first leaf node via the first plane, or (ii) that the third leaf node can communicate with the first leaf node via the second plane.
14. The system of claim 9 , the operations further comprising, wherein causing the second leaf node to communicate using the second communication path to the first leaf node using the second plane includes:
sending a control-plane instruction to the second leaf node that instructs the second leaf node to refrain from load balancing communication flows to the first leaf node through the first plane.
15. The system of claim 9 , the operations further comprising:
identifying a third leaf node having a third communication path to the first leaf node that is affected by the failure in the first plane;
determining that the second leaf node has a first number of communication flows with the first leaf node;
determining that the third leaf node has a second number of communication flows with the first leaf node, wherein the first number of communication flows is greater than the second number of communication flows; and
notifying the second leaf node of the failure in the first plane prior to notifying the third leaf node based at least in part on the first number of communication flows being greater than the second number of communication flows.
16. One or more non-transitory computer-readable media storing computer-executable instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
identifying, by an orchestration component, a failure in a first plane of a multi-plane network;
determining, by the orchestration component, that communication paths to a first leaf node are affected by the failure in the first plane;
identifying a first set of nodes in the multi-plane network that have first communication paths that are affected by the failure, wherein a second set of nodes in the multi-plane network have second communication paths that are unaffected by the failure;
notifying the first set of nodes indicating the failure while refraining from notifying the second set of nodes about the failure;
identifying a second leaf node having a first communication path to the first leaf node that is affected by the failure in the first plane; and
causing the second leaf node to communicate with the first leaf node using a second communication path in a second plane of the multi-plane network.
17. The one or more non-transitory computer-readable media of claim 16 , the operations further comprising:
receiving, at the orchestration component, telemetry data from a plurality of devices in the multi-plane network;
analyzing the telemetry data to determine a plurality of communication paths between a plurality of pairs of leaf nodes in the multi-plane network; and
generating mappings between individual pairs of leaf nodes of the plurality of pairs of leaf nodes based on the individual pairs of leaf nodes having at least one communication path,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying, from the mappings, that the second leaf node is mapped to the first leaf node.
18. The one or more non-transitory computer-readable media of claim 16 , the operations further comprising:
receiving, from the second leaf node, a request to monitor the first communication path to the first leaf node; and
storing a mapping between the second leaf node and the first leaf node based at least in part on receiving the request,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying the mapping.
19. The one or more non-transitory computer-readable media of claim 16 , the operations further comprising:
receiving a request to establish a secure tunnel between the first leaf node and the second leaf node;
causing the first leaf node and the second leaf node to establish the secure tunnel over the first communication path;
storing a mapping between the second leaf node and the first leaf node based at least in part on the secure tunnel being established between the second leaf node and the first leaf node,
wherein identifying the second leaf node having the first communication path to the first leaf node includes identifying the mapping.
20. The one or more non-transitory computer-readable media of claim 16 , the operations further comprising:
receiving, at the orchestration component, telemetry data representing communications between a plurality of devices in the multi-plane network over a period of time;
analyzing the telemetry data to determine that a third leaf node has previously established a third communication path with the first leaf node in the first plane during the period of time, wherein the third leaf node is not currently communicating with the first leaf node;
sending an instruction to the third leaf node indicating that the third leaf node either (i) cannot reach the first leaf node via the first plane, or (ii) that the third leaf node can communicate with the first leaf node via the second plane.