Managing a hierarchical network
The following description is directed to managing a hierarchical network including one or more network zones. In one example, a method of managing a hierarchical network includes collecting network state from respective devices of a network zone; using the collected network state to build an aggregated topology of the network zone; and transmitting the aggregated topology of the network zone to a traffic engineering service.
1. A method of managing a network zone comprising a plurality of network devices arranged in a leaf and spine architecture, the method comprising:
collecting operational network state from respective devices of the network zone, the operational network state comprising connectivity, link state, and capacity of the respective devices;
based on the collected operational network state, using a network zone controller to compile an aggregated topology including a link state section comprising a port group associated with a given leaf of the network zone, the port group having an effective capacity and a plurality of logical ports, wherein:
a respective logical port of the plurality of logical ports is associated with a client connected to the given leaf of the network zone and having a usable capacity, wherein the given leaf comprises a plurality of interconnected network devices, and
the effective capacity of the port group represents an internal capacity of the port group and the effective capacity of the port group is not greater than a number of in-service columns multiplied by a minimum capacity of the in-service columns, a respective in-service column being an active connection from the leaf to the spine;
providing the compiled aggregated topology to a traffic engineering service, the traffic engineering service being different than the network zone controller;
receiving, from the traffic engineering service, a tunnel specification based on the aggregated topology; and
configuring the respective devices of the network zone based on the received tunnel specification.
2. The method of claim 1 , wherein the usable capacity of the respective logical port is based on a number of network links connecting the client and the given leaf.
3. The method of claim 1 , wherein configuring the respective devices of the network zone comprises programming one or more equal cost multi-path tables based on weights of the received tunnel specification.
4. The method of claim 1 , further comprising:
detecting a partial reduction in capacity associated with a column of the network zone; and
taking the column of the network zone out of service based on the partial reduction in capacity.
5. A computer-readable storage medium including instructions that upon execution cause a computer system to:
collect network state from respective devices of a network zone, the network zone comprising a plurality of leaves, a respective leaf of the plurality of leaves comprising a plurality of interconnected network devices from the respective devices of the network zone, the respective leaf being connected to one or more clients, and the network state comprising operational link state of the respective devices of the network zone;
use a network zone controller and the collected network state to build an aggregated topology descriptor of the network zone, the aggregated topology descriptor comprising:
an inter-network-zone effective capacity for each leaf of the plurality of leaves, wherein the inter-network-zone effective capacity for each leaf represents an internal capacity of the leaf and wherein the inter-network-zone effective capacity for each leaf is less than or equal to a number of in-service columns associated with the leaf multiplied by a minimum column bandwidth of the in-service columns;
an edge corresponding to each client connected to a respective leaf; and
an effective capacity corresponding to each edge; and
transmit the aggregated topology descriptor of the network zone to a traffic engineering service, the traffic engineering service being different than the network zone controller.
6. The computer-readable storage medium of claim 5 , wherein the effective capacity corresponding to the edge for each leaf is based on a number of network link connections between the leaf and the client.
7. The computer-readable storage medium of claim 5 , wherein the instructions, upon execution, further cause the computer system to:
receive a path specification based on the transmitted aggregated topology; and
configure the respective devices of the network zone based on the received path specification.
8. The computer-readable storage medium of claim 5 , wherein
configuring the respective devices of the network zone comprises programming one or more equal cost multi-path tables based on weights of the received path specification.
9. The computer-readable storage medium of claim 5 , wherein the instructions, upon execution, further cause the computer system to:
detect a reduction in capacity associated with a column of the network zone; and
take the column of the network zone out of service based on the reduction in capacity.
10. The computer-readable storage medium of claim 5 , wherein the client of the respective leaf is a service associated with the network zone.
11. The computer-readable storage medium of claim 5 , wherein the client of the respective leaf is a different network zone.
12. A system comprising: a traffic engineering service executing on a first group of one or more server computers within a particular network zone, the traffic engineering service for allocating bandwidth of a network; a network zone comprising a plurality of network devices for forwarding network traffic, the network devices arranged as a plurality of leaves connected by a plurality of columns, a respective leaf comprising a plurality of interconnected network devices from the plurality of network devices of the network zone, and the respective leaf being connected to one or more clients; and a network zone controller executing on a second group of one or more server computers, the second group of one or more server computers being different than the first group of one or more server computers, the network zone controller adapted to: generate an aggregated topology descriptor for the network zone based on collected network state of the respective devices of the network zone, the collected network state comprising operational link state of the respective devices of the network zone, the aggregated topology descriptor comprising: an inter-network-zone effective capacity for each leaf of the plurality of leaves, wherein the inter-network-zone effective capacity for each leaf is proportional to a number of in-service columns associated with the leaf multiplied by a minimum column bandwidth of the in-service columns and the inter-network-zone effective capacity is representative of an internal capacity for each leaf; an edge corresponding to each client connected to a respective leaf; and an effective capacity corresponding to each edge; and provide the aggregated topology descriptor for the network zone to the traffic engineering service.
13. The system of claim 12 , wherein the effective capacity corresponding to the edge for each leaf is based on a number of active network link connections between the leaf and the client.
14. The system of claim 12 , wherein each leaf of the network zone comprises a multi-tiered Clos-based network.
15. The system of claim 12 , wherein the network zone controller is further adapted to:
configure the respective devices of the network zone to forward the network traffic across the columns using equal cost multi-path routing.
16. The system of claim 12 , wherein the network zone controller is further adapted to:
configure the respective devices of the network zone to forward the network traffic between the one or more clients and the respective leaves using weighted cost multi-path routing.
17. The system of claim 12 , wherein the network zone controller is further adapted to:
detect a partial reduction in capacity associated with a given column of the plurality of columns of the network zone; and
take the given column of the network zone out of service based on the partial reduction in capacity.