Capacity engineering in distributed computing systems

1. A computer system comprising:

one or more processors; and

one or more hardware storage devices having stored thereon instructions that are executable by the one or more processors to configure the computer system to generate one or more sets of minimal link capacities and identify a particular set of minimal link capacities, including instructions that are executable to configure the computer system to perform at least the following:

obtain network traffic demand for a network, the network traffic demand describing a link capacity that the network needs to support;

obtain a plurality of different network topologies for the network, each network topology describing a different configuration of links and devices for the network;

obtain a plurality of failure scenarios for the network, including obtaining at least one failure scenario corresponding to each of the plurality of different network topologies, each failure scenario describing how the links and devices in the corresponding network topology can fail;

generate a group of sets of minimal link capacities for the plurality of different network topologies, including generating a set of optimized minimal link capacities for each of the plurality of different network topologies, by analyzing the network using an optimization technique that determines the set of optimized minimal link capacities for the network topology based on the network traffic demand for the network, the at least one failure scenario for the network topology, and the configuration of links and devices described for the network topology, each generated set of optimized minimal link capacities capable of meeting all network demands across all obtained failure scenarios for the network topology;

identify a particular set of minimal link capacities from the group of sets of minimal link capacities, based on one or more predetermined optimality criteria;

identify a particular network topology corresponding to the identified particular set of minimal link capacities; and

provide the particular set of optimized minimal link capacities and the corresponding particular network topology.

2. The computer system of claim 1 , wherein the optimization technique is a convex optimization technique.

3. The computer system of claim 1 , wherein the optimization technique is configured to simultaneously solve for each set of optimized minimal link capacities for the plurality of failure scenarios.

4. The computer system of claim 1 , wherein at least one failure scenario comprises a requirement on a number of link failures in one or more network topologies of the plurality of different network topologies.

5. The computer system of claim 1 , wherein at least one network topology of the plurality of different network topologies comprises a plurality of copies of at least one common design pattern.

6. The computer system of claim 1 , wherein a first predetermined optimality criterion of the one or more predetermined optimality criteria is a minimum net change in network resources and wherein identifying the particular set of minimal link capacities comprises at least comparing a net change in network resources of each network topology corresponding to each set of minimal link capacities of the group of sets of minimal link capacities.

7. The computer system of claim 6 , wherein identifying the particular set of minimal link capacities further comprises identifying the particular set of minimal link capacities as a first set of minimal link capacities that satisfies the first predetermined optimality criterion.

8. The computer system of claim 6 , wherein the minimum net change in network resources is based at least in part on a monetary cost associated with modifying the network based on a first network topology of the plurality of different network topologies.

9. The computer system of claim 1 , wherein each of the plurality of different network topologies is associated with one or more shared risk link groups that each comprises a group of a plurality of network links that would fail if one link of the group fails.

10. The computer system of claim 1 , wherein at least one failure scenario is associated with a service level agreement.

11. A computer system comprising:

one or more processors; and

one or more hardware storage devices having stored thereon instructions that are executable by the one or more processors to configure the computer system to perform network optimization and automatic network scaling in the network using global optimization across topology design and traffic engineering for the network, including instructions that are executable to configure the computer system to perform at least the following:

obtain network traffic demand for the network, the network traffic demand describing a link capacity that the network needs to support;

obtain a plurality of different network topologies for the network, each network topology describing a different configuration of links and devices for the network;

obtain a plurality of failure scenarios for the network, including obtaining at least one failure scenario corresponding to each of the plurality of different network topologies, each failure scenario describing how the links and devices in the corresponding network topology can fail;

generate a group of sets of minimal link capacities for the plurality of different network topologies, including generating a set of optimized minimal link capacities for each of the plurality of different network topologies, by analyzing the network using an optimization technique that determines the set of optimized minimal link capacities for the network topology based on the network traffic demand for the network, the at least one failure scenario for the network topology, and the configuration of links and devices described for the network topology, each generated set of optimized minimal link capacities capable of meeting all network demands across all obtained failure scenarios for the network topology;

identify a particular set of minimal link capacities from the group of sets of minimal link capacities, based on one or more predetermined optimality criteria;

identify a particular network topology corresponding to the identified particular set of minimal link capacities; and

create an updated network by updating the network based on the particular set of minimal link capacities.

12. The computer system of claim 11 , wherein the optimization technique is a convex optimization technique.

13. The computer system of claim 11 , wherein the optimization technique is configured to simultaneously solve for each set of minimal link capacities for the plurality of failure scenarios.

14. The computer system of claim 11 , wherein creating the updated network comprises at least requisitioning new network resources defined in the network topology corresponding to the particular set of minimal link capacities.

15. The computer system of claim 11 , wherein each of the plurality of different network topologies is associated with one or more shared risk link groups that each comprises a group of a plurality of network links that would fail if one link of the group fails.

16. The computer system of claim 11 , wherein at least one failure scenario is associated with a service level agreement.

17. A method, implemented at a computer system that includes one or more processors, for network optimization and automatic network scaling in a network using global optimization across topology design and traffic engineering for the network, the method comprising:

obtaining network traffic demand for the network, the network traffic demand describing a link capacity that the network needs to support;

obtaining a plurality of different network topologies for the network, each network topology describing a different configuration of links and devices for the network;

obtaining a plurality of failure scenarios for the network, including obtaining at least one failure scenario corresponding to each of the plurality of different network topologies, each failure scenario describing how the links and devices in the corresponding network topology can fail;

generating a group of sets of minimal link capacities for the plurality of different network topologies, including generating a set of optimized minimal link capacities for each of the plurality of different network topologies, by analyzing the network using an optimization technique that determines the set of optimized minimal link capacities for the network topology based on the network traffic demand for the network, the at least one failure scenario for the network topology, and the configuration of links and devices described for the network topology, each generated set of optimized minimal link capacities capable of meeting all network demands across all obtained failure scenarios for the network topology;

identifying a particular set of minimal link capacities from the group of sets of minimal link capacities, based on one or more predetermined optimality criteria;

identifying a particular network topology corresponding to the identified particular set of minimal link capacities; and

creating an updated network by updating the network based on the particular set of minimal link capacities.

18. The method of claim 17 , wherein the optimization technique is a convex optimization technique.

19. The method of claim 17 , wherein the optimization technique is configured to simultaneously solve for each set of minimal link capacities for the plurality of failure scenarios.

20. The method of claim 17 , wherein creating the updated network comprises at least requisitioning new network resources defined in the network topology corresponding to the particular set of minimal link capacities.