Granted Patent B2

US 12,395,377 · App. 18/516,543 · Granted Aug 19, 2025

Inventors: David Paul Wragg (London, GB); Ólafur GuÐmundsson (Chevy Chase, MD); Lorenz Mathias Bauer (London, GB); Arthur Fabre (London, GB); Marek Przemyslaw Majkowski (Warsaw, PL)

Assignee: CLOUDFLARE, INC.

H04L67/1008H04L45/22H04L47/122H04L67/1001

Abstract

Claims (85)

1. A method of load balancing traffic between a plurality of Points of Presence (PoPs) of a cloud computing infrastructure that provides a cloud computing service to one or more users, the method comprising:

receiving feedback data from the plurality of PoPs, the feedback data related to states of compute servers in the plurality of PoPs;

identifying, based on the feedback data, a first POP of the plurality of PoPs from which traffic needs to be forwarded to another POP;

identifying, based on the feedback data, a second POP of the plurality of PoPs towards which traffic can be forwarded from another POP;

determining an inter-POP forwarding policy defining an amount of first traffic received by the first POP to forward to the second POP;

transmitting the inter-POP forwarding policy to the plurality of PoPs;

receiving, at the first POP, the inter-PoP forwarding policy;

converting, at the first PoP, the inter-POP forwarding policy to forwarding table entries of a forwarding table used in the first POP, wherein each forwarding table entry includes a packet identifier that corresponds to a packet flow and a network address associated with the second POP, wherein the forwarding table entries are configured to direct a proportion of packet flows that align with the defined amount of traffic received by the first POP to be forwarded to the second POP;

receiving, at the first POP, a first packet;

determining a first packet identifier of the first packet that identifies a first packet flow;

determining, based on a first forwarding table entry of the forwarding table for the first packet identifier, a first network address associated with the second PoP; and

transmitting the first packet to the first network address associated with the second PoP.

2. The method of claim 1 , wherein identifying, based on the feedback data, the first POP of the plurality of PoPs from which the traffic needs to be forwarded to another PoP comprises:

determining that a CPU load of at least one compute server at the first POP exceeds a threshold value.

3. The method of claim 1 , wherein identifying, based on the feedback data, the second POP of the of the plurality of PoPs towards which the traffic can be forwarded from another POP comprises:

determining that a CPU load of one or more compute servers at the second PoP is lower than a threshold value.

4. The method of claim 1 , wherein determining the inter-PoP forwarding policy defining the amount of the first traffic received by the first POP to forward to the second POP further comprises:

specifying a first amount of the first traffic of a first traffic type to be forwarded from the first POP to the second POP, wherein the first traffic type is associated with a first cloud computing service provided by the cloud computing infrastructure; and

specifying a second amount of the first traffic of a second traffic type to be forwarded from the first POP to the second POP, wherein the second traffic type is associated with a second cloud computing service provided by the cloud computing infrastructure.

5. The method of claim 4 , further comprising:

receiving updated feedback data from the plurality of PoPs; and

modifying the specified first amount of the first traffic of the first traffic type to be forwarded from the first POP to the second POP and the specified second amount of the first traffic of the second traffic type to be forwarded from the first POP to the second PoP based on the updated feedback data.

6. The method of claim 1 , further comprising:

identifying, based on the feedback data, a third POP of the plurality of PoPs towards which traffic can be forwarded from another PoP;

specifying a first amount of the first traffic of a first traffic type to be forwarded from the first POP to the second POP, wherein the first traffic type is associated with a first cloud computing service provided by the cloud computing infrastructure; and

specifying a second amount of the first traffic of a second traffic type to be forwarded from the first POP to the third POP, wherein the second traffic type is associated with a second cloud computing service provided by the cloud computing infrastructure.

7. The method of claim 1 , wherein the feedback data for a compute server in the plurality of PoPs includes one or more of a first parameter indicating a CPU load of the compute server, a second parameter measuring a Quality of Service provided by a service implemented on the compute server, and a measure of latency of the service.

8. A system, comprising:

a first non-transitory machine-readable storage medium that provides instructions that, if executed by a first processor of a central controller of a cloud computing infrastructure that provides a cloud computing service to one or more users, will cause said central controller to perform operations including:

receiving feedback data from a plurality of Point of Presences (PoPs) of the cloud computing infrastructure, the feedback data related to states of compute servers in the plurality of PoPs;

identifying, based on the feedback data, a first POP of the plurality of PoPs from which traffic needs to be forwarded to another POP;

identifying, based on the feedback data, a second POP of the plurality of PoPs towards which traffic can be forwarded from another POP;

determining an inter-POP forwarding policy defining an amount of first traffic received by the first POP to forward to the second POP; and

transmitting the inter-PoP forwarding policy to the plurality of PoPs;

a second non-transitory machine-readable storage medium that provides instructions that, if executed by one or more second processors of the first PoP will cause operations to be performed including:

receiving, at the first POP, the inter-POP forwarding policy;

converting, at the first POP, the inter-PoP forwarding policy to forwarding table entries of a forwarding table used in the first POP, wherein each forwarding table entry includes a packet identifier that corresponds to a packet flow and a network address associated with the second POP, wherein the forwarding table entries are configured to direct a proportion of packet flows that align with the defined amount of traffic received by the first POP to be forwarded to the second PoP;

receiving, at the first POP, a first packet;

determining a first packet identifier of the first packet that identifies a first packet flow;

determining, based on a first forwarding table entry of the forwarding table for the first packet identifier, a first network address associated with the second POP; and

transmitting the first packet to the first network address associated with the second POP.

9. The system of claim 8 , wherein identifying, based on the feedback data, the first POP of the plurality of PoPs from which the traffic needs to be forwarded to another POP further causes the central controller to perform operations including:

determining that a CPU load of at least one compute server at the first POP exceeds a threshold value.

10. The system of claim 8 , wherein identifying, based on the feedback data, the second POP of the plurality of PoPs towards which the traffic can be forwarded from another POP further causes the central controller to perform operations including:

determining that a CPU load of one or more compute servers at the second POP is lower than a threshold value.

11. The system of claim 8 , wherein determining the inter-POP forwarding policy defining the amount of the first traffic received by the first PoP to forward to the second POP further causes the central controller to perform operations including:

specifying a first amount of the first traffic of a first traffic type to be forwarded from the first PoP to the second POP, wherein the first traffic type is associated with a first cloud computing service provided by the cloud computing infrastructure; and

specifying a second amount of the first traffic of a second traffic type to be forwarded from the first POP to the second POP, wherein the second traffic type is associated with a second cloud computing service provided by the cloud computing infrastructure.

12. The system of claim 11 , wherein the instructions further cause the central controller to perform operations including:

receiving updated feedback data from the plurality of PoPs; and

modifying the specified first amount of the first traffic of the first traffic type to be forwarded from the first POP to the second PoP and the specified second amount of the first traffic of the second traffic type to be forwarded from the first PoP to the second PoP based on the updated feedback data.

13. The system of claim 8 , wherein the instructions further cause the central controller to perform operations including:

identifying, based on the feedback data, a third POP of the plurality of PoPs towards which traffic can be forwarded from another POP;

specifying a first amount of the first traffic of a first traffic type to be forwarded from the first PoP to the second POP, wherein the first traffic type is associated with a first cloud computing service provided by the cloud computing infrastructure; and

specifying a second amount of the first traffic of a second traffic type to be forwarded from the first POP to the third POP, wherein the second traffic type is associated with a second cloud computing service provided by the cloud computing infrastructure.

14. The system of claim 8 , wherein the feedback data for a compute server in the plurality of PoPs includes one or more of a first parameter indicating a CPU load of the compute server, a second parameter measuring a Quality of Service provided by a service implemented on the compute server, and a measure of latency of the service.

15. A system, comprising:

a central controller of a cloud computing infrastructure that includes a first processor; and a non-transitory machine-readable storage medium that provides instructions that, if executed by the first processor, will cause the central controller to perform operations including:

receiving feedback data from a plurality of Point of Presences (PoPs) of the cloud computing infrastructure, the feedback data related to states of compute servers in the plurality of PoPs;

identifying, based on the feedback data, a first POP of the plurality of PoPs from which traffic needs to be forwarded to another PoP;

identifying, based on the feedback data, a second POP of the plurality of PoPs towards which traffic can be forwarded from another POP;

determining an inter-POP forwarding policy defining an amount of first traffic received by the first POP to forward to the second POP; and

transmitting the inter-POP forwarding policy to the plurality of PoPs;

the first POP that includes a plurality of compute servers that each include a second non-transitory machine readable storage medium that provides instructions that, if executed by one or more second processors of the first POP will cause operations to be performed including:

receiving, at the first POP, the inter-PoP forwarding policy;

converting, at the first POP, the inter-POP forwarding policy to forwarding table entries of a forwarding table used in the first POP, wherein each forwarding table entry includes a packet identifier that corresponds to a packet flow and a network address associated with the second PoP, wherein the forwarding table entries are configured to direct a proportion of packet flows that align with the defined amount of traffic received by the first POP to be forwarded to the second POP;

receiving, at the first POP, a first packet;

determining a first packet identifier of the first packet that identifies a first packet flow;

determining, based on a first forwarding table entry of the forwarding table for the first packet identifier, a first network address associated with the second POP; and

transmitting the first packet to the first network address associated with the second PoP.

16. The system of claim 15 , wherein identifying, based on the feedback data, the first POP of the plurality of PoPs from which the traffic needs to be forwarded to another POP causes the central controller to perform operations comprising:

determining that a CPU load of at least one compute server at the first POP exceeds a threshold value.

17. The system of claim 15 , wherein identifying, based on the feedback data, the second POP of the plurality of PoPs towards which the traffic can be forwarded from another POP causes the central controller to perform operations comprising:

determining that a CPU load of one or more compute servers at the second POP is lower than a threshold value.

18. The system of claim 15 , wherein determining the inter-PoP forwarding policy defining the amount of the first traffic received by the first POP to forward to the second POP causes the central controller to perform operations comprising:

specifying a first amount of the first traffic of a first traffic type to be forwarded from the first PoP to the second POP, wherein the first traffic type is associated with a first cloud computing service provided by the cloud computing infrastructure; and

specifying a second amount of the first traffic of a second traffic type to be forwarded from the first POP to the second POP, wherein the second traffic type is associated with a second cloud computing service provided by the cloud computing infrastructure.

19. The system of claim 18 , wherein the operations performed by the central controller further comprise:

receiving updated feedback data from the plurality of PoPs; and

modifying the specified first amount of the first traffic of the first traffic type to be forwarded from the first POP to the second POP and the specified second amount of the first traffic of the second traffic type to be forwarded from the first POP to the second PoP based on the updated feedback data.

20. The system of claim 15 , wherein the operations performed by the central controller further comprise:

identifying, based on the feedback data, a third POP of the plurality of PoPs towards which traffic can be forwarded from another POP;

specifying a first amount of the first traffic of a first traffic type to be forwarded from the first PoP to the second POP, wherein the first traffic type is associated with a first cloud computing service provided by the cloud computing infrastructure; and

specifying a second amount of the first traffic of a second traffic type to be forwarded from the first POP to the third POP, wherein the second traffic type is associated with a second cloud computing service provided by the cloud computing infrastructure.

21. The system of claim 15 , wherein the feedback data for a compute server in the plurality of PoPs includes one or more of a first parameter indicating a CPU load of the compute server, a second parameter measuring a Quality of Service provided by a service implemented on the compute server, and a measure of latency of the service.

Assignments (2)

Continuity (3)

References Cited (28)