Methods and systems for making effective use of system resources

Patent metadata

Field	Value
Publication number	US-10192169-B2
Application number	US-201514703682-A
Country	US
Kind code	B2
Filing date	May 4, 2015
Priority date	Dec 10, 2010
Publication date	Jan 29, 2019
Grant date	Jan 29, 2019

Abstract

Methods and systems for making effective use of system resources. A plurality of requests for access to a resource are received. Each request has an associated group of features. The group of features for each request is analyzed to collect observations about the plurality of requests. A function to predict an outcome of a subsequent request is generated based on the observations. Resources are allocated to service the subsequent request based on the function.

First claim

What is claimed is:

1. A method comprising:
receiving a plurality of requests for access to a resource with a training module of a on-demand services environment provided by one or more computing devices, each request having an associated group of features to be mapped to labels within the on-demand services environment;
generating a statistical model with the training module within the on-demand services environment, wherein the statistical model provides a prediction of resources required by a subsequent request based on features of the corresponding subsequent requests, the subsequent request received from a remote computing device;
allocating resources within the on-demand services environment to service the subsequent request based on the prediction;
utilizing the allocated resources within the on-demand services environment with the one or more computing devices to generate a result; and
transmitting the result to the remote computing device.

2. The method of claim 1 , wherein prediction of resources required by the subsequent request comprises an execution time and a number of rows.

3. The method of claim 1 wherein allocating resources within the database comprises distribution of requests among application servers within the on-demand services environment.

4. The method of claim 1 wherein the plurality of requests comprises database queries.

5. The method of claim 4 wherein the group of features comprises:
a number of nested joins in the request;
a number of hash joins in the request; and
a sum of cardinalities for hash joins in the request.

6. The method of claim 4 wherein the group of features comprises:
a cardinality estimate of a main table;
a storage count of the main table; and
a selectivity estimate of the main table.

7. The method of claim 4 wherein the group of features comprises a number of secondary queries in the request.

8. The method of claim 4 wherein the group of features comprises:
an application server processor state; and
an application server memory state.

9. The method of claim 4 wherein the group of features comprises:
a database processor state; and
a number of tables used.

10. A non-transitory computer readable medium having stored there on instructions that, when executed, cause one or more processors to:
receive a plurality of requests for access to a resource with a training module of a database system provided by one or more computing devices, each request having an associated group of features to be mapped to labels in a database environment;
generate a statistical model with the training module within the database environment, wherein the statistical model provides a prediction of resources required by a subsequent request based on features of the corresponding subsequent requests, the subsequent request received from a remote computing device;
allocate resources within the database environment to service the subsequent request based on the prediction;
utilize the allocated resources within the database environment with the one or more computing devices to generate a result; and
transmit the result to the remote computing device.

11. The non-transitory computer-readable medium of claim 10 , wherein prediction of resources required by the subsequent request comprises an execution time and a number of rows.

12. The non-transitory computer-readable medium of claim 10 wherein allocating resources within the database comprises distribution of requests among application servers within the on-demand services environment.

13. The non-transitory computer-readable medium of claim 10 wherein the plurality of requests comprises database queries.

14. The non-transitory computer-readable medium of claim 13 wherein the group of features comprises:
a number of nested joins in the request;
a number of hash joins in the request; and
a sum of cardinalities for hash joins in the request.

15. The non-transitory computer-readable medium of claim 13 wherein the group of features comprises:
a cardinality estimate of a main table;
a storage count of the main table; and
a selectivity estimate of the main table.

16. The non-transitory computer-readable medium of claim 13 wherein the group of features comprises a number of secondary queries in the request.

17. The non-transitory computer-readable medium of claim 13 wherein the group of features comprises:
an application server processor state; and
an application server memory state.

18. The non-transitory computer-readable medium of claim 13 wherein the group of features comprises:
a database processor state; and
a number of tables used.

19. An on-demand services system comprising one or more server computing devices communicatively coupled to receive requests from one or more user computing devices, the system to receive a plurality of requests for access to a resource with a training module of the system, each request having an associated group of features to be mapped to labels in a on-demand services system, to generate a statistical model with the training module within the database environment, wherein the statistical model provides a prediction of resources required by a subsequent request based on features of the corresponding subsequent requests, the subsequent request received from one of the user computing devices, to allocate resources within the database environment to service the subsequent request based on the prediction, to utilize the allocated resources within the database environment with the one or more computing devices to generate a result, and to transmit the result to the remote computing device.

20. The system of claim 19 , wherein prediction of resources required by the subsequent request comprises an execution time and a number of rows.

21. The system of claim 19 wherein allocating resources within the database comprises distribution of requests among application servers within the on-demand services environment.

22. The system of claim 19 wherein the plurality of requests comprises database queries.

23. The system of claim 22 wherein the group of features comprises:
a number of nested joins in the request;
a number of hash joins in the request; and
a sum of cardinalities for hash joins in the request.

24. The system of claim 22 wherein the group of features comprises:
a cardinality estimate of a main table;
a storage count of the main table; and
a selectivity estimate of the main table.

25. The system of claim 22 wherein the group of features comprises a number of secondary queries in the request.

26. The system of claim 22 wherein the group of features comprises:
an application server processor state; and
an application server memory state.

27. The system of claim 22 wherein the group of features comprises:
a database processor state; and
a number of tables used.

Assignees

• Salesforce Com Inc

Inventors

• Gusev Andrey
• Yang Ronald
• Hansma Scott
• Collins Jesse
• Arbizu Alan

Classifications

• H04L43/0876

  Network utilisation, e.g. volume of load or congestion level · CPC title

• G06F16/2455

  Query execution · CPC title

• G06F9/5011

  the resources being hardware resources other than CPUs, Servers and Terminals · CPC title

• H04L47/125

  by balancing the load, e.g. traffic engineering · CPC title

• G06F2209/5019

  Workload prediction · CPC title