Dependency handling in an on-demand network code execution system

• US 10,725,752 B1
• Filed: 02/13/2018
• Issued: 07/28/2020
• Est. Priority Date: 02/13/2018
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Wagner, Timothy Allen, Moffatt, Bryan Nicholas, Brooker, Marc John, Rodrigues dos Santos, Robison, Jayakar, Niranjan
• Primary Examiner(s): Do, Chat C
• Assistant Examiner(s): Paulino, Lenin
• Application Number:
  US15/895,449
• Time in Patent Office: 896 Days
• CPC Class Codes: G06F 16/183 Provision of network file services by network file servers, e.g. by using NFS, CIFS network file access protocols H04L67/1097
• CPC Class Codes: G06F 2009/45575 Starting, stopping, suspending or resuming virtual machine instances
• CPC Class Codes: G06F 8/433 Dependency analysis; Data or control flow analysis
• CPC Class Codes: G06F 8/4441 Reducing the execution time required by the program code
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects

Abstract

Systems and methods are described for managing dependencies among code in an on-demand code execution system in a manner that enables efficient execution of the code. A user may generate a task on the system by submitting code, which may reference additional code on the system. The system can determine a dependency graph for the submitted code, and generate multiple representations of the dependency graph, storing dependency code objects in different numbers of files. On receiving a request to provision an execution environment for the task, the system can select an appropriate representation based on an expected reuse of the files within different representations. For example, a representation of fewer files can be selected where expected future use is low, minimizing retrieval time but also potentially limiting reuse of code objects within the representation.

Claims

1. A system comprising:

a physical data store storing;

a first package representing as a single file a set of code objects depended on by at least first user-submitted code executable on an on-demand code execution system; and

a collection of packages representing the set of code objects as multiple files;

a computing device configured with executable instructions to;

receive a request for the set of code objects depended on by at least first user-submitted code executable;

make a selection between returning information identifying the first package or information identifying the collection of packages in response to the request, wherein making the selection comprises;

determining, for the collection of packages, a threshold use rate for incurring additional time required to transfer the set of code objects as the collection of packages to a computing device within the on-demand code execution system relative to time required to transfer the set of code objects as the first package to the computing device, wherein the threshold use rate is determined based at least partly on a number of files within the collection of packages;

obtaining an expected future use rate on the on-demand code execution system for the collection of packages, wherein the expected future use rate indicates at least one expected future use of a file, within the multiple files, by an execution of second user-submitted executable code on the on-demand code execution system, the second user-submitted code being distinct from the first user-submitted code;

determining whether the expected future use rate for the collection of packages at least meets the threshold use rate for incurring the additional time required to transfer the collection of packages to the computing device, wherein if the expected future use rate at least meets the threshold, the computing device is configured to return information identifying the collection of packages in response to the request, and wherein if the expected future use rate does not meet the threshold, the computing device is configured to return information identifying the first packages in response to the request; and

in response to determining that the expected future use rate for the collection of packages at least meets the threshold use rate, selecting to return the information identifying the collection of packages in response to the request; and

return, as a response to the request, the information identifying the collection of packages, wherein the response to the request causes the computing device within the on-demand code execution system to retrieve individual files within the collection of packages for use during execution of the first user-submitted executable code.

2. The system of claim 1, wherein the set of code objects includes to at least one of a code library, a framework, a software development kit (SDKs), a driver, or a runtime.

3. The system of claim 1, wherein the expected future use rate for the collection of packages represents a number of expected future uses for the collection over a period of time.

4. The system of claim 3, wherein the expected future use rate for the collection of packages represents a sum of expected future uses for individual files within the collection of packages.

5. A computer-implemented method comprising:

receiving a request for a set of code objects depended on by first user-submitted code executable on an on-demand code execution system;

identifying a first set of files representing the set of code objects and a second set of files representing the set of code objects, wherein a number of files within the second set of files is greater than a number of files within the first set of files;

obtaining, for the second set of files, a threshold use value for incurring additional time to transfer the set of code objects as the second set of files to a computing device within the on-demand code execution system relative to time required to transfer the set of code objects as the first set of files to the computing device, wherein the threshold use value is determined based at least partly on files within the second set of files;

making a selection between returning information identifying the first set of files or the second set of files in response to the request, wherein making the selection comprises;

determining that an expected future use value on the on-demand code execution system for the second set of files satisfies the threshold use value for incurring the additional time required to transfer the second set of files to the computing device, wherein the expected future use value indicates at least one expected future use of a file, within the second set of files, by an execution of second user-submitted executable code on the on-demand code execution system, the second user-submitted code being distinct from the first user-submitted code; and

in response to determining that the expected future use value for the second set of files satisfies the threshold use value, selecting to return the information identifying the second set of files in response to the request;

returning, in response to the request, the information identifying the second set of files;

retrieving the second set of files using the information and

provisioning the set of code objects from the second set of files into an execution environment for the first user-submitted code.

6. The computer-implemented method of claim 5, wherein the first set of files comprises a single file.

7. The computer-implemented method of claim 5, wherein provisioning the set of code objects from the second set of files into the execution environment for the first user-submitted code comprises generating a file system for the execution environment and extracting the set of code objects from the second set of files into the file system.

8. The computer-implemented method of claim 7, wherein the file system is a union mount file system.

9. The computer-implemented method of claim 5, wherein the expected future use on the on-demand code execution system for the second set of files is an expected future use by a computing device from which the request was received.

10. The computer-implemented method of claim 5 further comprising determining the expected future use on the on-demand code execution system for the second set of files at least partly by determining a past number of executions on the on-demand code execution system of a plurality of sets of user-submitted code that depend on code objects included within an individual file within the set of files.

11. The computer-implemented method of claim 5, wherein determining that the expected future use value on the on-demand code execution system for the second set of files satisfies the threshold use value comprises determining that the expected future use value for the second set of files exceeds the threshold use value by at least an expected future use value on the on-demand code execution system for the first set of files.

12. Non-transitory computer readable media including computer-executable instructions that, when executed on an on-demand code execution system, causes the on-demand code execution system to:

receive a request for a set of code objects depended on by first user-submitted code executable on an on-demand code execution system;

identify a first set of files representing the set of code objects and a second set of files representing the set of code objects, wherein a number of files within the second set of files is greater than a number of files within the first set of files;

obtain, for the second set of files, a threshold use value for incurring additional time to transfer the set of code objects as the second set of files to a computing device within the on-demand code execution system relative to time to transfer the set of code objects as the first set of files to the computing device, wherein the threshold use value is determined based at least partly on files within the second set of files;

make a selection between returning information identifying the first set of files or the second set of files in response to the request, wherein making the selection comprises;

determining that an expected future use value on the on-demand code execution system for the second set of files exceeds the threshold use value for incurring the additional time required to transfer the second set of files to the computing device, wherein the expected future use value indicates at least one expected future use of a file, within the second set of files, by an execution of second user-submitted executable code on the on-demand code execution system, the second user-submitted code being distinct from the first user-submitted code; and

in response to determining that the expected future use value for the second set of files satisfies the threshold use value, selecting to return the information identifying the second set of files in response to the request;

return, in response to the request, information identifying the second set of files;

retrieve the second set of files using the information; and

provision the set of code objects from the second set of files into an execution environment for the first user-submitted code.

13. The non-transitory computer-readable media of claim 12, wherein the expected future use value for the second set of files represents a sum of expected future uses for individual files within the second set of files.

14. The non-transitory computer-readable media of claim 12, wherein the second set of files are retrieved from a network storage service external to the on-demand code execution system.

15. The non-transitory computer-readable media of claim 12, wherein provisioning the set of code objects from the second set of files into the execution environment for the first user-submitted code comprises generating a file system for the execution environment and extracting the set of code objects from the second set of files into the file system.

16. The non-transitory computer-readable media of claim 12, wherein the expected future use on the on-demand code execution system for the second set of files is an expected future use by a collection of computing devices associated with a shared cache.

17. The non-transitory computer-readable media of claim 12 wherein the instructions further cause the on-demand code execution system to determine the expected future use on the on-demand code execution system for the second set of files at least partly by determining a past number of executions on the on-demand code execution system of a plurality of sets of user-submitted code that depend on code objects included within an individual file within the set of files.

18. The non-transitory computer-readable media of claim 12 wherein the instructions further cause the on-demand code execution system to determine that the expected future use value on the on-demand code execution system for the second set of files satisfies the threshold use value at least partly by determining that the expected future usage value for the second set of files exceeds the threshold use value by at least an expected future use value on the on-demand code execution system for the first set of files.

19. The non-transitory computer-readable media of claim 12, wherein the threshold use value is determined based at least partly on a number of files within the second set of files, a format of the files within the second set of files, or attributes of code within the files of the second set of files.