Granted Patent B1

US 12,639,100 · App. 18/958,570 · Granted May 26, 2026

Inventors: Prachetaa Raghavan (Sammamish, WA); Ajit Yagaty (Sammamish, WA); Roger Hoover (San Mateo, CA); Godwin Pang (Pasadena, CA); Sahil Gandhi (Morgan Hill, CA); Thomas Snyder (New York, NY); Sunil Patil (Los Altos, CA); Deyu Jiao (Seattle, WA); Rajesh RC (Sunnyvale, CA)

Assignee: Confluent, Inc.

G06F9/45558G06F2009/4557

Abstract

Claims (47)

1 . A system, comprising:

at least one processor; and

at least one memory including instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising:

providing a first plane at a controller, the first plane being associated with a plurality of regional agents associated with a plurality of object groups where each regional agent of the plurality of regional agents comprises an object group of the plurality of object groups associated therewith;

associating a second plane with an object group of each of the plurality of object groups, the object group having a pod group where:

each pod group of the plurality of object groups has an availability zone divided into sub-availability zones, a number of sub-availability zones in the availability zone dynamically varying based on processing changes of nodes associated with the object groups;

each of the availability zones are associated with the second plane, the second plane employing virtual machines having a plurality of data planes that mediate and control network communication; and

a first sub-availability zone of the availability zones being configured to substitute functionality of a second sub-availability zone of the sub-availability zones, the first sub-availability zone and the second sub-availability zone sharing a data plane of the plurality of data planes where the first sub-availability zone has first functionality and the second sub-availability zone has second functionality;

adding additional sub-availability zones to the availability zone by further dividing the availability zone based on the processing changes;

migrating the second functionality of the second sub-availability zone to the first sub-availability zone based on a failure of the second sub-availability zone;

monitoring an operational state of the second sub-availability zone to detect when the second sub-availability zone becomes functional after the failure; and

based on the second sub-availability zone becoming functional after the failure, migrating the second functionality back to the second sub-availability zone, where the shared data plane minimizes data loss during the migrations.

2 . The system of claim 1 , wherein the instructions further cause the system to perform operations comprising adding a third sub-availability zone in response to a change in demand.

3 . The system of claim 2 , wherein each of the first sub-availability zone, the second sub-availability zone, and the third sub-availability zone are siloed from each other.

4 . The system of claim 3 , wherein the pod group has a plurality of pods and the instructions further cause the system to perform operations comprising scaling the pod group to include additional nodes in response to a change in demand.

5 . The system of claim 1 , wherein the pod group has a plurality of pods and the instructions further cause the system to perform operations comprising scaling the pod group to include additional nodes in response to a change in demand.

6 . The system of claim 1 , wherein the first plane is a control plane and the second plane is a data plane.

7 . The system of claim 1 , wherein the first plane is an abstraction layer and the second plane is a data plane.

8 . A method comprising:

providing a first plane at a controller, the first plane being associated with a plurality of regional agents associated with a plurality of object groups where each regional agent of the plurality of regional agents comprises an object group of the plurality of object groups associated therewith;

associating a second plane with an object group of each of the plurality of object groups, the object group having a pod group where:

each pod group of the plurality of object groups has an availability zone divided into sub-availability zones, a number of sub-availability zones in the availability zone dynamically varying based on processing changes of nodes associated with the object groups;

each of the availability zones are associated with the second plane, the second plane employing virtual machines having a plurality of data planes that mediate and control network communication; and

a first sub-availability zone of the availability zones being configured to substitute functionality of a second sub-availability zone of the sub-availability zones, the first sub-availability zone and the second sub-availability zone sharing a data plane of the plurality of data planes where the first sub-availability zone has first functionality and the second sub-availability zone has second functionality;

adding additional sub-availability zones to the availability zone by further dividing the availability zone based on the processing changes;

migrating the second functionality of the second sub-availability zone to the first sub-availability zone based on a failure of the second sub-availability zone;

monitoring an operational state of the second sub-availability zone to detect when the second sub-availability zone becomes functional after the failure; and

based on the second sub-availability zone becoming functional after the failure, migrating the second functionality back to the second sub-availability zone, where the shared data plane minimizes data loss during the migrations.

9 . The method of claim 8 , comprising:

adding a third sub-availability zone in response to a change in demand and each of the first sub-availability zone, the second sub-availability zone, and the third sub-availability zone being siloed from each other; and

scaling the pod group to include additional nodes in response to a change in demand.

10 . The method of claim 8 , wherein the pod group has a plurality of pods and the method further comprises scaling the pod group to include additional nodes in response to a change in demand.

11 . The method of claim 8 , wherein the first plane is one of a control plane or an abstraction layer and the second plane is a data plane.

12 . A machine-storage medium having instructions embodied thereon, the instructions executable by at least one hardware processor to perform operations comprising:

providing a first plane at a controller, the first plane being associated with a plurality of regional agents associated with a plurality of object groups where each regional agent of the plurality of regional agents comprises an object group of the plurality of object groups associated therewith;

associating a second plane with an object group of each of the plurality of object groups, the object group having a pod group where:

each pod group of the plurality of object groups has an availability zone divided into sub-availability zones, a number of sub-availability zones in the availability zone dynamically varying based on processing changes of nodes associated with the object groups;

each of the availability zones are associated with the second plane, the second plane employing virtual machines having a plurality of data planes that mediate and control network communication; and

a first sub-availability zone of the availability zones being configured to substitute functionality of a second sub-availability zone of the sub-availability zones, the first sub-availability zone and the second sub-availability zone sharing a data plane of the plurality of data planes where the first sub-availability zone has first functionality and the second sub-availability zone has second functionality;

adding additional sub-availability zones to the availability zone by further dividing the availability zone based on the processing changes;

migrating the second functionality of the second sub-availability zone to the first sub-availability zone based on a failure of the second sub-availability zone;

monitoring an operational state of the second sub-availability zone to detect when the second sub-availability zone becomes functional after the failure; and

based on the second sub-availability zone becoming functional after the failure, migrating the second functionality back to the second sub-availability zone, where the shared data plane minimizes data loss during the migrations.

13 . The machine-storage medium of claim 12 , wherein the operations comprise:

adding a third sub-availability zone in response to a change in demand and each of the first sub-availability zone, the second sub-availability zone, and the third sub-availability zone being siloed from each other; and

scaling the pod group to include additional nodes in response to a change in demand.

14 . The machine-storage medium of claim 12 , wherein the pod group has a plurality of pods and the operations further comprise scaling the pod group to include additional nodes in response to a change in demand.

Assignments (2)

References Cited (8)