Automated profiling of resource usage

• US 10,542,079 B2
• Filed: 06/27/2018
• Issued: 01/21/2020
• Est. Priority Date: 09/20/2012
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Marr, Michael David, Klein, Matthew D.
• Primary Examiner(s): Scott, Randy A
• Application Number:
  US16/020,477
• Publication Number:
  US 20180332107A1
• Time in Patent Office: 573 Days
• Field of Search: 709220, 709224, 709225, 709226, 718 1, 718104
• CPC Class Codes: G06F 9/5044 considering hardware capabilities
• CPC Class Codes: H04L 41/0806 for initial configuration or provisioning, e.g. plug-and-play
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30

Abstract

Operating profiles for consumers of computing resources may be automatically determined based on an analysis of actual resource usage measurements and other operating metrics. Measurements may be taken while a consumer, such as a virtual machine instance, uses computing resources, such as those provided by a host. A profile may be dynamically determined based on those measurements. Profiles may be generalized such that groups of consumers with similar usage profiles are associated with a single profile. Assignment decisions may be made based on the profiles, and computing resources may be reallocated or oversubscribed if the profiles indicate that the consumers are unlikely to fully utilize the resources reserved for them. Oversubscribed resources may be monitored, and consumers may be transferred to different resource providers if contention for resources is too high.

Claims

1. A system for managing shared computing resources, the system comprising:

one or more processors; and

a computer-readable memory storing executable instructions that, when executed by the one or more processors, configure the system to at least;

obtain a first resource profile associated with a first virtual machine, wherein a first instance of the first virtual machine is executed on a first computing device of a plurality of computing devices;

obtain a second resource profile associated with a second virtual machine, wherein a second instance of the second virtual machine is executed on the first computing device;

determine, based at least in part on the first resource profile, a first expected usage amount of a computing resource at a particular time;

determine, based at least in part on the second resource profile, a second expected usage amount of the computing resource at the particular time;

calculate a total expected usage amount, the total expected usage amount comprising a sum of the first expected usage amount and the second expected usage amount;

determine that the total expected usage amount is greater than an available amount of the computing resource on the first computing device at the particular time;

in response to determining that the total expected usage amount is greater than the available amount;

determine that an existing virtual machine instance executed on the second computing device is using or is expected to use less than all of a quantity of the computing resource reserved for the existing virtual machine instance on the second computing device;

select the second computing device from the plurality of computing devices based at least in part on determining that the existing virtual machine instance executed on the second computing device is using less than all of a quantity of the computing resource reserved for the existing virtual machine instance on the second computing device; and

continue execution of the first instance of the first virtual machine on the first computing device and transfer execution of the second instance to a second computing device of the plurality of computing devices.

2. The system of claim 1, wherein the computing resource comprises at least one of processing resources, storage, memory, network, hard disk, or electrical power.

3. The system of claim 1, wherein the first expected usage amount and the second expected usage amount relate to memory capacity, processing capacity, storage, network bandwidth, network latency, position within a network topology, instruction set, or variance of a performance metric.

4. The system of claim 1, wherein to select the second computing device, the computer-readable memory further stores executable instructions that, when executed by the one or more processors, configure the system to at least determine that a characteristic associated with the computing device corresponds to the second expected usage amount.

5. The system of claim 4, wherein to select the second computing device, the computer-readable memory further stores executable instructions that, when executed by the one or more processors, configure the system to at least determine that the second expected usage amount is less than an available amount of the computing resource on the second computing device at the particular time.

6. A computer-implemented method for managing shared computing resources, the computer-implemented method comprising:

obtaining a first resource profile associated with a first virtual machine, wherein a first instance of the first virtual machine is executed on a first computing device of a plurality of computing devices;

obtaining a second resource profile associated with a second virtual machine, wherein a second instance of the second virtual machine is executed on the first computing device;

calculating a total expected usage amount, the total expected usage amount comprising a sum of a first expected usage amount based at least in part on the first resource profile and a second expected usage amount based at least in part on the second resource profile;

determining that the total expected usage amount is greater than an available amount of a computing resource on the first computing device at a particular time;

in response to determining that the total expected usage amount is greater than the available amount;

determining that an existing virtual machine instance executed on the second computing device is using or is expected to use less than all of a quantity of the computing resource reserved for the existing virtual machine instance on the second computing device;

selecting the second computing device based at least in part on determining that the existing virtual machine instance executed on the second computing device is using less than all of a quantity of the computing resource reserved for the existing virtual machine instance on the second computing device; and

continuing execution of the first instance of the first virtual machine on the first computing device and transferring execution of the second instance to a second computing device of the plurality of computing devices.

7. The computer-implemented method of claim 6, further comprising:

selecting the second computing device based at least in part on a number of virtual machine instances previously hosted on the second host computing device and associated with a common customer.

8. The computer-implemented method of claim 6, wherein the first usage amount and the second usage amount are associated with operating metrics that vary over the lifecycle of the instance of the virtual machine.

9. The computer-implemented method of claim 8, wherein at least one of the operating metrics relates to processing unit utilization, memory utilization, network utilization, hard disk utilization, or electrical power utilization.

10. The computer-implemented method of claim 6, wherein the first expected usage amount is determined based at least in part on a generalized configuration associated with the first virtual machine.

11. The computer-implemented method of claim 6, wherein the first expected usage amount is determined based at least in part on the generalized configuration associated with the first virtual machine and data received from a customer associated with the first virtual machine.

12. The computer-implemented method of claim 6, wherein at least one of the first expected usage amount and the second expected usage amount is determined by at least one historical measure of the usage of the amount of the computing resource.

13. A non-transitory, computer-readable medium having stored thereon computer-executable software instructions configured to cause one or more processors associated with a system for managing shared computing resources to perform operations comprising:

calculating a total expected usage amount for a computing resource, the total expected usage amount comprising a sum of a first expected usage amount associated with a first virtual machine executed on a first computing device and a second expected usage amount associated with a second virtual machine executed on the first computing device;

determining that the total expected usage amount is greater than an available amount of the computing resource on the first computing device at a particular time;

in response to determining that the total expected usage amount is greater than the available amount;

determining that an existing virtual machine instance executed on the second computing device is using or is expected to use less than all of a quantity of the computing resource reserved for the existing virtual machine instance on the second computing device;

selecting the second computing device based at least in part on determining that the existing virtual machine instance executed on the second computing device is using less than all of a quantity of the computing resource reserved for the existing virtual machine instance on the second computing device; and

continuing execution of the first virtual machine on the first computing device and transferring execution of the second, virtual machine to a second computing device.

14. The non-transitory, computer-readable medium of claim 13, wherein the first expected usage amount and the second expected usage amount are determined based at least in part on a service level agreement with a customer associated with a minimal operating parameter for the computing resource.

15. The non-transitory, computer-readable medium of claim 13, wherein the first resource profile comprises a customer-specific resource profile, and wherein the first expected usage amount is based at least M part on a historical operating metric associated with initialization or use of the first virtual machine associated with the customer-specific profile.

16. The non-transitory, computer-readable medium of claim 13, wherein at least one of the first expected usage amount and the second expected usage amount comprises a median, standard deviation, or usage histogram of a historical operating metric.