Systems and methods providing optimization data

• US 9,292,336 B1
• Filed: 01/22/2014
• Issued: 03/22/2016
• Est. Priority Date: 01/22/2014
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Ramalingam, Harsha, Jadhav, Manoj Bhagwan, Mannepalli, Srinivasa Susheel Kumar, Webster, Scott Cameron
• Primary Examiner(s): Chen, Xi D
• Application Number:
  US14/161,412
• Time in Patent Office: 790 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202

Abstract

Service fleets made up of many pieces of computer hardware may perform computational tasks. Described herein are systems and methods for optimizing costs associated with the computer hardware. In one example, an optimization system indicates an amount of hardware needed such that the service fleet operates at an optimal cost. The amount of hardware may be determined based on hardware cost metric data generated for a service fleet.

Claims

1. A method comprising:

generating, via at least one of one or more computing devices, throughput data

representative of quantities of transactions processed over a period of time, wherein the transactions are processed by a service operating on hardware, wherein the service comprises a set of instructions that execute to perform computational tasks;

accessing expense data representative of an expense associated with the hardware;

generating, via at least one of the one or more computing devices, hardware cost metric data indicative of a first cost associated with a performance of the hardware, wherein the hardware cost metric data is generated based on the expense data and the throughput data;

accessing threshold data indicative of a second cost, the second cost comprising an optimal cost associated with the performance of the hardware based on at least one of the hardware cost metric data reaching a threshold capacity;

accessing prediction data indicative of a predicted quantity of transactions to be processed by the service;

generating predicted hardware cost metric data indicative of a third cost comprising a predicted cost associated with the performance of the hardware, wherein;

the predicted hardware cost metric data is generated based on the hardware cost metric data and the prediction data; and

generating, via at least one of the one or more computing devices, optimization data indicative of an amount of the hardware that changes the first cost to the second cost, wherein the optimization data is generated based on the predicted hardware cost metric data and the threshold data.

2. The method of claim 1, further comprising:

accessing proposed service data representative of a proposed service, the proposed service being different from the service operating on the hardware;

determining that in response to the proposed service being executed on the hardware, at least one or more performance metrics will be equal to or less than a threshold value;

generating procurement data indicating a need to reduce or terminate an order of additional hardware previously ordered; and

provide the procurement data to a supplier computing device associated with a supplier that supplies the additional hardware.

3. The method of claim 1, further comprising:

accessing proposed service data representative of a proposed service comprising a set of instructions that would execute on the hardware, the proposed service being different from the service operating on the hardware;

determining that in response to the proposed service being executed on the hardware, at least one or more performance metrics will be equal to or more than a threshold value;

generating alert data comprising information about the occurrence the one or more performance metrics being equal or more than the threshold value; and

providing the alert data to client computing device associated with an administrative user.

4. The method of claim 1, further comprising accessing test data representative of sample quantities of transactions, wherein the threshold data is generated based on the expense data and the test data.

5. The method of claim 1, further comprising:

generating first metric data comprising a first value representing information about a processor capacity utilization associated with the service;

generating second metric data comprising a second value representing information about a memory capacity utilization associated with the service; and

generating third metric data comprising a third value representing information about an input/output operations per second utilization associated with the service.

6. The method of claim 5, further comprising:

determining that at least one of the first value, the second value, or the third value is equal to or greater than a threshold value; and

in response to the at least one of the first value, the second value, or the third value being equal to or greater than a threshold value, determining the hardware cost metric data.

7. The method of claim 1, wherein the optimization of the performance of the hardware includes changing from the first cost to the second cost.

8. The method of claim 1, wherein the threshold data is generated based on a benchmark test configured to determine an amount of sample transactions which causes the at least one of a plurality of metric values to reach the threshold capacity.

9. The method of claim 1 wherein the predicted quantity of transactions is predicted based on at least one of:

a historical analysis of quantities of transactions processed by the service;

an introduction of a new product;

or

a planned reduction in price for one or more products.

10. The method of claim 1, further comprising:

generating procurement data based on the optimization data, the procurement data indicative of an amount of additional hardware to be ordered for the service; and

providing the procurement data to a supplier that supplies additional hardware.

11. The method of claim 1, further comprising accessing one or more log files comprising recorded information about activities performed by the service, wherein the throughput data is generated by analyzing the one or more log files.

12. The method of claim 1, wherein the first cost is determined by dividing the expense associated with the hardware by the peak quantity of transactions occurring over the period of time.

13. The method of claim 1, further comprising:

determining one or more business drivers associated with one or more of a plurality of application programming interfaces, wherein the service includes the plurality of application programming interfaces; and

determining prediction data based on the business driver.

15. The system of claim 1, wherein the first cost is generated by dividing the expense associated with the hardware by the peak quantity of transactions occurring over the period of time.

16. The system of claim 1, wherein the server computing device is further configured to:

generate procurement data based on the optimization data, the procurement data indicative of an amount of additional hardware to be ordered for the service; and

provide the procurement data to a computing device executing an order processing service associated with a supplier of additional hardware.

14. A system comprising:

a server computing device configured including code that when executed causes the server computing device to;

generate throughput data representative of quantities of transactions processed over a period of time, wherein;

the transactions comprise application programming interface calls; and

the transactions are processed by a service operating on hardware, the service comprising a set of instructions that execute to perform computational tasks;

generate peak data representative of a peak quantity of transactions occurring over the period of time, wherein the peak data is generated based on the throughput data;

access expense data representative of an expense associated with the hardware;

generate hardware cost metric data indicative of a first cost associated with a performance of the hardware, wherein the hardware cost metric data is generated based on the expense data and the peak data;

access prediction data representative of a predicted quantity of transactions to be processed by the service;

generate predicted hardware cost metric data indicative of a second cost comprising a predicted cost associated with the performance of the hardware, wherein the predicted hardware cost metric data is generated based on the hardware cost metric data and the prediction data;

access test data representative of sample quantities of transactions and at least one of the hardware cost metric data reaching a threshold capacity;

generate threshold data indicative of a third cost comprising an optimal cost associated with the performance of the hardware, wherein the threshold data is generated based on the expense data and the test data;

generate business driver data indicating one or more dominant application interface calls comprising one or more of the application interface calls which have more of an impact on the performance of the hardware when compared to other application interface calls;

generate scaling formula data used to determine an amount of hardware consumed for the one or more dominant application interface calls; and

generate optimization data representative of an amount of the hardware that changes the first cost to the third cost, wherein the optimization data is generated based on the predicted hardware cost metric data, the threshold data, the business driver data, and the scaling formula data.

17. A non-transitory computer readable medium storing instructions, which when executed by a processor, cause the processor to perform actions comprising:

generating throughput data representative of quantities of transactions processed over a period of time, wherein the transactions are processed by a service operating on hardware;

accessing expense data representative of an expense associated with the hardware;

generating hardware cost metric data indicative of a first cost associated with a performance of the hardware;

accessing threshold data indicative of a second cost, the second cost comprising an optimal cost associated with the performance of the hardware based on determining when at least one of the hardware cost metric data reaches a threshold capacity;

accessing prediction data indicative of a predicted quantity of transactions to be processed by the service;

generating predicted hardware cost metric data indicative of a third cost comprising a predicted cost associated with the performance of the hardware, wherein;

the predicted hardware cost metric data is generated based on the hardware cost metric data and the prediction data; and

generating optimization data comprising information indicating an amount of the hardware that changes the first cost to the second cost, wherein optimization data is generated based on the hardware cost metric data, the threshold data, and the prediction data.

18. The computer readable medium of claim 17, wherein the threshold data is generated based on a benchmark test configured to determine an amount of sample transactions which causes the at least one of a plurality of metric values to reach the threshold capacity.

19. The computer readable medium of claim 17, the actions further comprising generating peak data representative of a peak quantity of transactions occurring over the period of time, wherein:

the throughput data is generated based on the peak data; and

the hardware cost metric data is generated based on the expense data and the peak data.

20. The computer readable medium of claim 17, wherein the predicted quantity of transactions is predicted based on a historical analysis of quantities of transactions processed by the service for the same time of year.