Enhancing API service schemes

• US 10,642,585 B1
• Filed: 03/16/2017
• Issued: 05/05/2020
• Est. Priority Date: 10/13/2014
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Srivastava, Kumar
• Primary Examiner(s): Anya, Charles E
• Assistant Examiner(s): Hoang, Phuong N
• Application Number:
  US15/461,012
• Time in Patent Office: 1,146 Days
• CPC Class Codes: G06F 11/3612 by runtime analysis performance monitoring G06F11/3466
• CPC Class Codes: G06F 8/443 Optimisation
• CPC Class Codes: G06F 9/451 Execution arrangements for user interfaces
• CPC Class Codes: G06F 9/48 Program initiating; Program switching, e.g. by interrupt
• CPC Class Codes: G06F 9/542 Event management; Broadcasting; Multicasting; Notifications
• CPC Class Codes: H04L 67/00 Network arrangements or protocols for supporting network services or applications user-to-user messaging H04L51/00; network arrangements, protocols or services for supporting real-time applications in data packet communications networks H04L65/00

Abstract

Embodiments of enhancing API servicing schemes are disclosed. A set of proxy code associated with servicing the API requests of an application is analyzed during development time to determine if any portion can be modified to improve the service scheme, prior to executing the set of proxy code in a runtime. After the set of proxy code is deployed in a runtime environment, performance data associated with the current service scheme is collected and periodically used to determine whether the service scheme for the API requests of the application is to be updated.

Claims

1. A computer-implemented method, comprising:

determining a current service scheme that specifies a sequential order that each service call out in a set of service call outs for serviced API requests for an application will be sent out to respective service servers for processing;

collecting performance data for the serviced API requests for the application, wherein the API requests were serviced based on the current service scheme that specifies the sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to respective service servers for processing;

determining whether a new service scheme for servicing the application at runtime is to be determined for the application based at least in part on a new service scheme condition, wherein the new service scheme specifies a different sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing;

in response to determining that a new service scheme for servicing the application at runtime is to be determined based at least in part on the new service scheme condition, determining a new service scheme that specifies another sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application, wherein in the another sequential order a first service call out is ordered before a second service call out based on a failure score of the first service call out as indicated by the performance data being more than a failure score of the second service call out as indicated by the performance data;

receiving an indication of approval for the new service scheme; and

servicing a received API request associated with the application using the new service scheme that was determined based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application.

2. The computer-implemented method of claim 1, wherein collecting performance data for the serviced API requests for the application comprises:

receiving an API request associated with the application;

obtaining data associated with servicing the API request using the current service scheme; and

storing, in memory, properties associated with the API request, the properties comprising one or more of;

identifying information associated with the application with which the API request was associated, a time at which the API request was sent to or received at a proxy server, and a date at which the API request was sent to or received at the proxy server.

3. The computer-implemented method of claim 2, wherein collecting performance data for the serviced API requests for the application further comprises:

storing, in memory, properties associated with a requestor associated with the API request, the properties comprising one or more of;

a type of device from which the API request originated, a type of operating system running on the device from which the API request originated, or identifying information associated with a user that has logged into the application.

4. The computer-implemented method of claim 2, wherein collecting performance data for the serviced API requests for the application further comprises:

storing, in memory, properties associated with each of the service callouts, which is sent to its respective target service, made for the API request according to the current service scheme, the properties comprising one or more of;

a type associated with the target service, a target data located at a target URL, an amount of time after which a response to the service callout was received from the target service, whether an error was received from the target service, a type of error that was received from the target service, current CPU utilization of the target service, current memory utilization of the target service, or current I/O utilization of the target service.

5. The computer-implemented method of claim 4, wherein collecting performance data for the serviced API requests for the application further comprises:

storing, in memory, a result of servicing the API request,

wherein a result of servicing the API request is successful if a response to the API request was successfully generated based on responses to the service callouts, and

wherein a result of servicing the API request is unsuccessful if either a response to the API request was not successfully generated based on responses to the service callouts or any one of the service callouts failed to obtain requested data from the target service.

6. The computer-implemented method of claim 1, wherein the new service scheme condition comprises one or more of:

at least one property of the current service scheme matches a suboptimal scheme rule, a threshold amount of performance data has been collected, or a threshold of failures of a serviced API request has been satisfied.

7. The computer-implemented method of claim 1, wherein the method is implemented using proxy code executed at a proxy server, and wherein the new service scheme is determined based at least in part on one or more of:

modifying at least a portion of a current set of proxy code, modifying parameters associated with the proxy server, modifying logic associated with the proxy server, selecting to use a different service scheme option that is included in the set of proxy code, selecting a different execution path that is included in the current set of proxy code, or causing the current set of proxy code to offload processing to a third party entity.

8. The method of claim 1, wherein servicing a received API request associated with the application using the new service scheme that was determined based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application comprises:

receiving a particular API request;

sending out each of the service call outs required to service the particular API request in the different order in accordance with the new service scheme;

receiving responses for each of the service call outs based on the different order; and

providing a response to the particular API request based on the responses for each of the service call outs.

9. The method of claim 1, wherein the new service scheme specifies, for each API request that is serviced using the new service scheme, an ordered set of the service call outs required to be sent out to service the respective API request.

10. The method of claim 1, wherein determining a new service scheme that specifies another sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application comprises:

determining the new service scheme to specify that two service call outs that were specified to be sent out sequentially in the current service scheme to be instead sent out in parallel.

11. The method of claim 1, wherein determining a new service scheme that specifies another sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application comprises:

in response to determining that a particular service call has a historical failure rate that satisfies a threshold, determining the new service scheme to specify to send out the particular service call multiple times.

12. The method of claim 1, wherein determining a new service scheme that specifies another sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application comprises:

determining the another sequential order based on ordering each of the service call outs in order by failure scores of each of the service call outs, wherein a failure score of a service call out reflects a historical response failure rate from the service call.

13. A system comprising one or more processors and one or more storage devices storing instructions that are operable, when executed by the one or more processors, to cause the one or more processors to perform operations comprising:

determining a current service scheme that specifies a sequential order that each service call out in a set of service call outs for serviced API requests for an application will be sent out to respective service servers for processing;

collecting performance data for the serviced API requests for the application, wherein the API requests were serviced based on the current service scheme that specifies the sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to respective service servers for processing;

determining whether a new service scheme for servicing the application at runtime is to be determined for the application based at least in part on a new service scheme condition, wherein the new service scheme specifies a different sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing;

in response to determining that a new service scheme for servicing the application at runtime is to be determined based at least in part on the new service scheme condition, determining a new service scheme that specifies another sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application, wherein in the another sequential order a first service call out is ordered before a second service call out based on a failure score of the first service call out as indicated by the performance data being more than a failure score of the second service call out as indicated by the performance data;

receiving an indication of approval for the new service scheme; and

servicing a received API request associated with the application using the new service scheme that was determined based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application.

14. The system of claim 13, wherein collecting performance data for the serviced API requests for the application comprises:

receiving an API request associated with the application;

obtaining data associated with servicing the API request using the current service scheme; and

storing, in memory, properties associated with the API request, the properties comprising one or more of;

identifying information associated with the application with which the API request was associated, a time at which the API request was sent to or received at a proxy server, and a date at which the API request was sent to or received at the proxy server.

15. The system of claim 14, wherein collecting performance data for the serviced API requests for the application further comprises:

storing, in memory, properties associated with a requestor associated with the API request, the properties comprising one or more of;

a type of device from which the API request originated, a type of operating system running on the device from which the API request originated, or identifying information associated with a user that has logged into the application.

16. The system of claim 14, wherein collecting performance data for the serviced API requests for the application further comprises:

storing, in memory, properties associated with each of the service callouts, which is sent to its respective target service, made for the API request according to the current service scheme, the properties comprising one or more of;

a type associated with the target service, a target data located at a target URL, an amount of time after which a response to the service callout was received from the target service, whether an error was received from the target service, a type of error that was received from the target service, current CPU utilization of the target service, current memory utilization of the target service, or current I/O utilization of the target service.

17. The system of claim 16, wherein collecting performance data for the serviced API requests for the application further comprises:

storing, in memory, a result of servicing the API request,

wherein a result of servicing the API request is successful if a response to the API request was successfully generated based on responses to the service callouts, and

wherein a result of servicing the API request is unsuccessful if either a response to the API request was not successfully generated based on responses to the service callouts or any one of the service callouts failed to obtain requested data from the target service.

18. The system of claim 13, wherein the new service scheme condition comprises one or more of:

at least one property of the current service scheme matches a suboptimal scheme rule, a threshold amount of performance data has been collected, or a threshold of failures of a serviced API request has been satisfied.

19. The system of claim 13, wherein the instructions comprise proxy code executed at a proxy server, and wherein the new service scheme is determined based at least in part on one or more of:

modifying at least a portion of a current set of proxy code, modifying parameters associated with the proxy server, modifying logic associated with the proxy server, selecting to use a different service scheme option that is included in the set of proxy code, selecting a different execution path that is included in the current set of proxy code, or causing the current set of proxy code to offload processing to a third party entity.

20. A computer-readable storage device storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

determining a current service scheme that specifies a sequential order that each service call out in a set of service call outs for serviced API requests for an application will be sent out to respective service servers for processing;

collecting performance data for the serviced API requests for the application, wherein the API requests were serviced based on the current service scheme that specifies the sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to respective service servers for processing;

determining whether a new service scheme for servicing the application at runtime is to be determined for the application based at least in part on a new service scheme condition, wherein the new service scheme specifies a different sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing;

in response to determining that a new service scheme for servicing the application at runtime is to be determined based at least in part on the new service scheme condition, determining a new service scheme that specifies another sequential order that each of the service call outs in the set of service call outs for the API requests will be sent out to the respective service servers for processing based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application, wherein in the another sequential order a first service call out is ordered before a second service call out based on a failure score of the first service call out as indicated by the performance data being more than a failure score of the second service call out as indicated by the performance data;

receiving an indication of approval for the new service scheme; and

servicing a received API request associated with the application using the new service scheme that was determined based at least in part on at least a portion of the collected performance data for serviced API requests associated with the application.