Prioritization-based scaling of computing resources

• US 10,148,592 B1
• Filed: 06/29/2015
• Issued: 12/04/2018
• Est. Priority Date: 06/29/2015
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Pai, Derek Solomon, Truong, Alison Qing-Ning, Stone, Eric Samuel, Khalid, Ahmed Usman, Tang, Kai Fan
• Primary Examiner(s): Donabed, Ninos
• Application Number:
  US14/754,447
• Time in Patent Office: 1,254 Days
• CPC Class Codes: H04L 12/14 Charging , metering or billing arrangements for data wireline or wireless communications
• CPC Class Codes: H04L 12/1417 Advice of charge with threshold, e.g. user indicating maximum cost
• CPC Class Codes: H04L 41/0681 Configuration of triggering conditions
• CPC Class Codes: H04L 43/08 Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
• CPC Class Codes: H04L 43/0894 Packet rate
• CPC Class Codes: H04M 15/00 Arrangements for metering, time-control or time indication ; Metering, charging or billing arrangements for voice wireline or wireless communications, e.g. VoIP
• CPC Class Codes: H04M 15/83 Notification aspects
• CPC Class Codes: H04M 15/8351 before establishing a communication
• CPC Class Codes: H04M 15/852 Low balance or limit reached
• CPC Class Codes: H04M 15/88 Provision for limiting connection, or expenditure
• CPC Class Codes: H04M 15/883 linked escalation limits, i.e. establish, first or second limit

Abstract

Techniques are described for scaling a group of computing resources. A computing resource service receives a scaling policy for use in scaling the group of computing resources. The scaling policy specifies a target level for a resource utilization metric and magnitude-based changes to the group. The computing resource service receives information about a magnitude of a measurement for the resource utilization metric. The computing resource service determines, based at least in part on the scaling policy, one or more changes for the group and initiates the one or more changes in the group.

Claims

1. A computer-implemented method, comprising:

receiving, by one or more configured computing systems of a computing resource service, a prioritization policy from a customer of the computing resource service for use in scaling a group of computing nodes provided to the customer by the computing resource service, wherein each of a plurality of scaling policies specifies one or more associated changes to be performed to the group if criteria are satisfied for the scaling policy that include at least one resource utilization metric exceeding a threshold, and wherein the prioritization policy specifies priorities between the plurality of scaling policies;

receiving, by the one or more configured computing systems, information indicating multiple alarms that have been triggered corresponding to multiple scaling policies of the plurality, based on the criteria for the multiple scaling policies being satisfied by monitored values for one or more resource utilization metrics;

selecting, by the one or more configured computing systems and based at least in part on the specified priorities of the prioritization policy, a scaling policy from the multiple scaling policies corresponding to the triggered multiple alarms, wherein the selected scaling policy has a higher priority in the prioritization policy than one or more other unselected scaling policies of the multiple scaling policies;

determining, by the one or more configured computing systems, a change to the group from the one or more associated changes for the selected scaling policy; and

initiating, by the one or more configured computing systems and based on the higher priority of the selected scaling policy in the prioritization policy, the determined change in the group without initiating additional changes to the group that are associated with the one or more other unselected scaling policies of the multiple scaling policies.

2. The computer-implemented method of claim 1, further comprising:

monitoring multiple defined alarm triggers that are each associated with one of the one or more resource utilization metrics, wherein each of the alarm triggers specifies a target value; and

initiating, based at least in part on determining from the monitoring that each of the target values is reached, the multiple alarms.

3. The computer-implemented method of claim 2 wherein the multiple defined alarms triggers are defined by the customer.

4. The computer-implemented method of claim 2 wherein one of the defined alarm triggers is associated with a relative change from the specified target value for the one defined alarm trigger.

5. The computer-implemented method of claim 1 wherein the associated changes include a relative change in quantity to at least one computing node quantity associated with the group of computing nodes, and wherein the initiating of the determined change in the group includes changing the at least one computing node quantity by the relative change in quantity.

6. A non-transitory computer-readable medium having stored contents that, when executed, cause one or more computing systems of a computing resource service to at least:

receive, by the one or more computing systems, a prioritization policy from a customer of the computing resource service for use in scaling a group of computing nodes provided to the customer by the computing resource service, wherein each of a plurality of scaling policies specifies one or more associated changes to be performed to the group of computing nodes if criteria are satisfied for the scaling policy that include at least one resource utilization metric exceeding a threshold, and wherein the prioritization policy specifies priorities between the plurality of scaling policies;

receive, by the one or more computing systems, information indicating multiple scaling policies of the plurality that have been triggered, based on the criteria for the multiple scaling policies being satisfied by monitored values for one or more resource utilization metrics;

select, by the one or more computing systems and based at least in part on the specified priorities of the prioritization policy, a scaling policy from the multiple scaling policies, wherein the selected scaling policy has a higher priority in the prioritization policy than one or more other unselected scaling policies of the multiple scaling policies;

determine, by the one or more configured computing systems, a change to the group from the one or more associated changes for the selected scaling policy; and

initiate, by the one or more computing systems and based on the higher priority of the selected scaling policy in the prioritization policy, the determined change to the group without initiating additional changes to the group that are associated with the one or more other unselected scaling policies of the multiple scaling policies.

7. The non-transitory computer-readable medium of claim 6 wherein the group of computing nodes includes storage capabilities, wherein the determined change is a change to at least one of a specified amount or a specified type of the storage capabilities, and wherein the initiating of the determined change includes performing the determined change to the at least one of the specified amount or the specified type of the storage capabilities.

8. The non-transitory computer-readable medium of claim 6 wherein the stored contents include software instructions that, when executed, further cause the one or more computing systems to monitor the computing nodes of the group to determine when the computing nodes of the group cause measurements for resource utilization metrics associated with the multiple scaling policies to have magnitudes that exceed defined thresholds.

9. The non-transitory computer-readable medium of claim 6 wherein the determined change to the group is based on a difference between a specified absolute quantity and a current quantity of computing nodes in the group of computing nodes, and wherein the stored contents, when executed, further cause the one or more computing systems to change the current quantity of computing nodes in the group based at least in part on the difference.

10. The non-transitory computer-readable medium of claim 6 wherein the determined change to the group includes an indicated change to a type of one or more computing nodes of the group, and wherein the stored contents, when executed, further cause the one or more computing systems to perform the indicated change in the type of the one or more computing nodes of the group.

11. The non-transitory computer-readable medium of claim 6 wherein the determined change includes an indicated change to a geographical location of one or more computing nodes of the group, and wherein the stored contents, when executed, further cause the one or more computing systems to perform the indicated change to the geographical location of the one or more computing nodes.

12. The non-transitory computer-readable medium of claim 6 wherein the group further includes one or more storage resources, and wherein the stored contents, when executed, further cause the one or more computing systems to perform the determined change by modifying the one or more storage resources.

13. The non-transitory computer-readable medium of claim 6 wherein the prioritization policy specifies a priority for the selected scaling policy that is higher than one or more other priorities specified for other of the multiple scaling policies, and wherein the selecting of the scaling policy is based on that scaling policy having the higher specified priority.

14. The non-transitory computer-readable medium of claim 6 wherein the selected scaling policy specifies an associated change to the group as an outcome of a function applied to a resource utilization metric measurement, and wherein the initiating of the change to the group includes applying the outcome of the function as part of performing the change to the group.

15. A system comprising:

one or more processors of one or more computing systems; and

memory with stored computer executable instructions that, when executed by the one or more processors, cause the system to implement functionality of a computing resource service, including;

receiving a prioritization policy from a customer of the computing resource service for use in scaling a group of computing nodes provided to the customer by the computing resource service, wherein each of a plurality of scaling policies specifies one or more associated changes to be performed to the group of computing nodes if criteria are satisfied for the scaling policy that include at least one resource utilization metric exceeding a threshold, and wherein the prioritization policy specifies priorities between the plurality of scaling policies;

receiving information indicating multiple scaling policies of the plurality that have been triggered, based on the criteria for the multiple scaling policies being satisfied by monitored values for one or more resource utilization metrics;

selecting, based at least in part on the specified priorities of the prioritization policy, a scaling policy from the multiple scaling policies, wherein the selected scaling policy has a higher priority from the prioritization policy than one or more other unselected scaling policies of the multiple scaling policies;

determining a change to the group from the one or more associated changes for the selected scaling policy; and

initiating, based on the higher priority of the selected scaling policy in the prioritization policy, the determined change in the group without initiating additional changes to the group that are associated with the one or more other unselected scaling policies of the multiple scaling policies.

16. The system of claim 15 wherein:

the resource utilization metric associated with the selected scaling policy is based at least in part on one or more information sources external to the group of computing nodes; and

the memory further comprises computer executable instructions that, when executed by the one or more processors, cause the system to monitor information associated with the one or more external information sources.

17. The system of claim 15 wherein the one or more computing systems are part of the computing resource service.

18. The system of claim 15 wherein the computer executable instructions further cause the system to monitor the computing nodes of the group to receive the information about the multiple scaling policies being triggered.

19. The system of claim 15 wherein the prioritization policy specifies a highest order of priority associated with a first resource utilization metric of multiple resource utilization metrics used with the multiple scaling policies, and wherein the selecting of the scaling policy is based at least in part on the selected scaling policy using the first resource utilization metric.

20. The system of claim 19 wherein the initiating of the determined change in the group includes implementing multiple changes to the group selected from the one or more associated changes for the selected scaling policy.