Detecting impedance mismatches due to cross-service calls

• US 10,733,085 B1
• Filed: 02/05/2018
• Issued: 08/04/2020
• Est. Priority Date: 02/05/2018
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Wagner, Timothy Allen
• Primary Examiner(s): Leibovich, Yair
• Application Number:
  US15/888,946
• Time in Patent Office: 911 Days
• Field of Search: 714 381
• CPC Class Codes: G06F 11/3604 Software analysis for verifying properties of programs byte-code verification G06F9/44589
• CPC Class Codes: G06F 11/3612 by runtime analysis performance monitoring G06F11/3466
• CPC Class Codes: G06F 11/368 for test version control, e.g. updating test cases to a new software version
• CPC Class Codes: G06F 11/3688 for test execution, e.g. scheduling of test suites
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 9/547 Remote procedure calls [RPC]; Web services
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06N 5/046 Forward inferencing; Production systems

Abstract

Systems and methods are described for conducting static analysis of code invoking network-based services to identify, without requiring execution of the code, an impedance mismatch between an expected execution rate of the code and an invocation capacity of a service invoked within the code. A system is provided that may analyze code to detect both direct invocations of services, as well as indirect invocations caused by the direct invocations. The system can utilize information regarding directly or indirectly invoked services to determine whether an expected invocation rate of such services will exceed invocation capacity for the services. In some instances, the system can traverse a “call graph” of all services invoked by code either directly or indirectly to identify impedance mismatches through multiple levels of indirection.

Claims

1. A system to analyze code executable on an on-demand code execution system to detect an impedance mismatch with an invoked network-accessible service, the system comprising:

a hardware data store storing executable code submitted to the on-demand code execution system by a user device, the executable code including an invocation of a first network-accessible service; and

a computing device comprising a hardware processor, the computing device in communication with the hardware data store and configured to;

obtain an expected execution rate for the executable code on the on-demand code execution system;

conduct a static analysis of the executable code, independent of execution of the executable code, to identify, within the executable code, portions of the executable code that, when executed by the on-demand code execution system, would cause the on-demand code execution system to transmit a set of requests to invoke a first network-accessible service in communication with the on-demand code execution system via a network;

from at least the expected execution rate for the executable code and a number of requests within the set of requests to invoke the first network-accessible service that would be caused by execution of the executable code on the on-demand code execution system, determine an expected invocation rate for the first network-accessible service when executing the executable code on the on-demand code execution system at the expected execution rate;

identify an invocation capacity for the first network-accessible service;

compare the expected invocation rate for the first network-accessible service to the invocation capacity for the first network-accessible service to determine that the expected invocation rate exceeds the invocation capacity; and

transmit an indication that that the expected invocation rate exceeds the invocation capacity.

2. The system of claim 1, wherein the computing device further configured to:

determine that an expected output of the first network-accessible service is one or more invocations of a second network-accessible service;

determine, based at least partly on the expected invocation rate for the first network-accessible service and a number of invocations within the one or more invocations of the second network-accessible, an expected invocation rate for the second network-accessible service;

identify an invocation capacity for the second network-accessible service;

compare the expected invocation rate for the second network-accessible service to the invocation capacity for the second network-accessible service to determine that the expected invocation rate for the second network-accessible service exceeds the invocation capacity for the second network-accessible service; and

transmit an indication that that the expected invocation rate for the second network-accessible service exceeds the invocation capacity for the second network-accessible service.

3. The system of claim 1, wherein the expected execution rate for the executable code is obtained from the user device.

4. The system of claim 1, wherein the expected execution rate for the executable code is obtained based on a historical frequency with which triggering criteria for the executable code would have been satisfied.

5. The system of claim 1, wherein the invocation capacity for the first network-accessible service is based at least partly on at least one of capacity information obtained from querying the first network-accessible service or an amount of computing resources available to the first network-accessible service.

6. The system of claim 1, wherein the indication that that the expected invocation rate exceeds the invocation capacity includes an indication of the number of requests within the set of requests to invoke the first network-accessible service that would be caused by execution of the executable code and an identification of the portions of the executable code that, when executed by the on-demand code execution system, would cause the on-demand code execution system to transmit the set of requests.

7. A computer-implemented method to analyze code executable on an on-demand code execution system comprising:

obtaining from a user device executable code submitted for execution on the on-demand code execution system;

obtaining an expected execution rate for the executable code on the on-demand code execution system;

conducting a static analysis of the executable code, independent of execution of the executable code, to identify, within the executable code, portions of the executable code that, when executed by the on-demand code execution system, would cause the on-demand code execution system to transmit a set of requests to invoke a first network-accessible service in communication with the on-demand code execution system via a network;

from at least the expected execution rate for the executable code and a number of requests within the set of requests to invoke the first network-accessible service that would be caused by execution of the executable code on the on-demand code execution system, determining an expected invocation rate for the first network-accessible service when executing the executable code on the on-demand code execution system at the expected execution rate;

identifying an invocation capacity for the first network-accessible service;

comparing the expected invocation rate for the first network-accessible service to the invocation capacity for the first network-accessible service to determine that the expected invocation rate exceeds the invocation capacity; and

transmitting an indication that that the expected invocation rate exceeds the invocation capacity.

8. The computer-implemented method of claim 7, wherein identifying the portions of the executable code comprises identifying a code segment within the executable code that is associated with the first network-accessible service.

9. The computer-implemented method of claim 8, wherein identifying the portions of the executable code further comprises referencing information mapping the code segment to the first network-accessible service.

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

determining that an expected output of the first network-accessible service corresponds to one or more invocations of a second network-accessible service;

determining, based at least partly on the expected invocation rate for the first network-accessible service and a number of invocations within the one or more invocations of the second network-accessible service, an expected invocation rate for the second network-accessible service;

identifying an invocation capacity for the second network-accessible service;

comparing the expected invocation rate for the second network-accessible service to the invocation capacity for the second network-accessible service to determine that the expected invocation rate for the second network-accessible service exceeds the invocation capacity for the second network-accessible service; and

transmitting an indication that that the expected invocation rate for the second network-accessible service exceeds the invocation capacity for the second network-accessible service.

11. The computer-implemented method of claim 10, wherein the expected invocation rate for the second network-accessible service is a multiplicative combination of the expected invocation rate for the first network-accessible service and the number of invocations within the one or more invocations of the second network-accessible service.

12. The computer-implemented method of claim 10, wherein the second network-accessible service is implemented as an execution of second executable code on an on-demand code execution system, and wherein determining that the expected output of the first network-accessible service corresponds to one or more invocations of the second network-accessible service comprises detecting within the second executable code, the one or more invocations.

13. The computer-implemented method of claim 10, wherein the second network-accessible service is implemented as an execution of second executable code on an on-demand code execution system, and wherein determining that the expected output of the first network-accessible service corresponds to one or more invocations of the second network-accessible service comprises determining that the expected output of the first network-accessible service satisfies criteria maintained at the on-demand code execution system for one or more executions of the second executable code.

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

determining, from at least the expected execution rate for the executable code and the invocations within the executable code of the first network-accessible service, an expected amount of data to be passed to the first network-accessible service;

determining a data capacity of the first network-accessible service;

determining that the expected amount of data to be passed to the first network-accessible service exceeds the data capacity for the first network-accessible service; and

transmitting an indication that that the expected amount of data to be passed to the first network-accessible service exceeds the data capacity for the first network-accessible service.

15. Non-transitory computer-readable media comprising executable instructions that, when executed on a computing system, cause the computing system to analyze code executable on an on-demand code execution system by at least:

obtaining from a user device executable code submitted for execution on an on-demand code execution system;

obtaining an expected execution rate for the executable code on the on-demand code execution system;

conducting a static analysis of the executable code, independent of execution of the executable code, to identify, within the executable code, portions of the executable code that, when executed by the on-demand code execution system, would cause the on-demand code execution system to transmit a set of requests to invoke a first network-accessible service in communication with the on-demand code execution system via a network;

from at least the expected execution rate for the executable code and a number of requests within the set of requests to invoke the first network-accessible service that would be caused by execution of the executable code on the on-demand code execution system, determining an expected invocation rate for the first network-accessible service when executing the executable code on the on-demand code execution system at the expected execution rate;

identifying an invocation capacity for the first network-accessible service;

comparing the expected invocation rate for the first network-accessible service to the invocation capacity for the first network-accessible service to determine that the expected invocation rate exceeds the invocation capacity; and

transmitting an indication that that the expected invocation rate exceeds the invocation capacity.

16. The non-transitory computer-readable media of claim 15, wherein the indication that that the expected invocation rate exceeds the invocation capacity includes an indication of the number of requests within the set of requests to invoke the first network-accessible service that would be caused by execution of the executable code and an identification of the portions of the executable code that, when executed by the on-demand code execution system, would cause the on-demand code execution system to transmit the set of requests.

17. The non-transitory computer-readable media of claim 15, wherein the executable instructions further cause the computing system to:

determine that an expected output of the first network-accessible service corresponds to one or more invocations of a second network-accessible service;

determine, based at least partly on the expected invocation rate for the first network-accessible service and a number of invocations within the one or more invocations of the second network-accessible service, an expected invocation rate for the second network-accessible service;

identify an invocation capacity for the second network-accessible service;

compare the expected invocation rate for the second network-accessible service to the invocation capacity for the second network-accessible service to determine that the expected invocation rate for the second network-accessible service exceeds the invocation capacity for the second network-accessible service; and

transmit to the user device an indication that that the expected invocation rate for the second network-accessible service exceeds the invocation capacity for the second network-accessible service.

18. The non-transitory computer-readable media of claim 17, wherein the expected invocation rate for the second network-accessible service is a multiplicative combination of the expected invocation rate for the first network-accessible service and the number of invocations within the one or more invocations of the second network-accessible service.

19. The non-transitory computer-readable media of claim 17, wherein the second network-accessible service is implemented as an execution of second executable code on an on-demand code execution system, and wherein determining that the expected output of the first network-accessible service corresponds to one or more invocations of the second network-accessible service comprises detecting within the second executable code, the one or more invocations.

20. The non-transitory computer-readable media of claim 15, wherein the executable instructions further cause the computing system to:

determine from at least the expected execution rate for the executable code and the invocations within the executable code of the first network-accessible service, an expected amount of data to be passed to the first network-accessible service;

determine a data capacity of the first network-accessible service;

determine that the expected amount of data to be passed to the first network-accessible service exceeds the data capacity for the first network-accessible service; and

transmit to the user device an indication that that the expected amount of data to be passed to the first network-accessible service exceeds the data capacity for the first network-accessible service.

21. The non-transitory computer-readable media of claim 20, wherein the expected amount of data to be passed to the first network-accessible service is based at least partly on one or more datatypes of parameters within the invocations of the first network-accessible service within the executable code.

22. The non-transitory computer-readable media of claim 15, wherein the executable instructions further cause the computing system to determine a modification to the expected execution rate for the executable code that would cause the expected invocation rate for the first network-accessible service to not exceed the invocation capacity for the first network-accessible service, and wherein the indication includes an indication of the modification.

23. The non-transitory computer-readable media of claim 22, wherein the indication of the modification is selectable to cause the modification to the expected execution rate.