Event-driven computing

• US 9,652,306 B1
• Filed: 09/29/2015
• Issued: 05/16/2017
• Est. Priority Date: 09/30/2014
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Nair, Ajay, Wisniewski, Scott Daniel, Wagner, Timothy Allen, Brooker, Marc John
• Primary Examiner(s): Dorais, Craig
• Application Number:
  US14/869,879
• Time in Patent Office: 595 Days
• Field of Search: None
• CPC Class Codes: G06F 11/3006 where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems multiprogramming arrangements G06F9/46; allocation of resources G06F9/50
• CPC Class Codes: G06F 11/301 where the computing system is a virtual computing platform, e.g. logically partitioned systems virtual machines G06F9/45533; logical partitioning of resources G06F9/5077
• CPC Class Codes: G06F 2009/45562 Creating, deleting, cloning virtual machine instances
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2009/45575 Starting, stopping, suspending or resuming virtual machine instances
• CPC Class Codes: G06F 2209/544 Remote
• CPC Class Codes: G06F 2209/548 Queue
• CPC Class Codes: G06F 9/45508 Runtime interpretation or emulation, e g. emulator loops, bytecode interpretation
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06F 9/542 Event management; Broadcasting; Multicasting; Notifications
• CPC Class Codes: G06F 9/547 Remote procedure calls [RPC]; Web services
• CPC Class Codes: H04L 41/06 Management of faults, events, alarms or notifications

Abstract

A service manages a plurality of virtual machine instances for low latency execution of user codes. The service can provide the capability to execute user code in response to events triggered on various event sources and initiate execution of other control functions to improve the code execution environment in response to detecting errors or unexpected execution results. The service may maintain or communicate with a separate storage area for storing code execution requests that were not successfully processed by the service. Requests stored in such a storage area may subsequently be re-processed by the service.

Claims

1. A system for processing event messages for requests to execute program code on one or more virtual machine instances, the system comprising:

a frontend computing system comprising one or more hardware computing devices executing specific computer-executable instructions, the frontend computing system in communication with (i) a virtual compute system adapted to provide compute capacity on a per-request basis and (ii) one or more other services, the frontend computing system configured to at least;

receive a plurality of code execution requests associated with a first program code, each code execution request including an identity of a user account, an identity of the first program code, and one or more first input parameters to the first program code, wherein the user account is associated with the first program code and a set of execution parameters used for executing program codes associated with the user account;

process the plurality of code execution requests at least by attempting to execute the first program code at a first time period using the one or more first input parameters and the set of execution parameters;

determine that one or more requests from the plurality of code execution requests were not processed successfully due to an error during execution of the first program code;

route the one or more requests to a separate storage area for additional processing, the separate storage area storing a plurality of requests that were not processed successfully by the frontend computing system;

determine whether a request stored in the separate storage area is to be processed again by the frontend computing system; and

in response to determining that the request is to be processed again by the frontend computing system, process the request at least by attempting to execute the first program code at a second time period subsequent to the first time period.

2. The system of claim 1, wherein the set of execution parameters comprises one or more levels of computing resources allocated to the program codes associated with the user account, the computing resources comprising one or more of CPU, memory, network, input/output (I/O), a maximum number of requests, or a maximum execution duration.

3. The system of claim 1, wherein the separate storage area comprises a dynamic dead letter queue (DLQ) configured to store code execution requests.

4. The system of claim 1, wherein the one or more other services comprise one of a storage service, a monitoring service, a logging service, a billing service, a data stream service, a messenger service, a database service, or a message queueing service.

5. A system, comprising:

a virtual compute system comprising one or more hardware computing devices executing specific computer-executable instructions and configured to at least maintain a plurality of pre-initialized virtual machine instances configured to execute program codes thereon;

a dynamic dead letter queue (DLQ) configured to store code execution requests; and

a frontend computing system comprising one or more hardware computing devices executing specific computer-executable instructions and configured to at least;

receive a plurality of code execution requests associated with the first program code, each code execution request including one or more first input parameters to the first program code, the first program code associated with configuration data;

process the plurality of code execution requests using the one or more first input parameters and the configuration data;

detect a problem in executing a subset of the plurality of code execution requests associated with the first program code;

without successfully executing the first program code for the subset of plurality of code execution requests, route the subset of the plurality of code execution requests to the dynamic DLQ;

determine that a request in the subset of the plurality of code execution requests in the dynamic DLQ is to be re-processed;

re-process the request at least by executing the first program code using the one or more first input parameters;

detect a trend in the processing of the plurality of code execution requests;

determine one or more second input parameters to a second program code based on the detected trend; and

cause the second program code to be executed on one of the plurality of pre-initialized virtual machine instances that is pre-initialized for executing the second program code using the one or more second input parameters.

6. The system of claim 5, wherein the second program code modifies the first program code by updating one or more parameters associated with the first program code, the one or more parameters comprising at least one of version information, permission information, runtime information, or library information.

7. The system of claim 5, wherein the second program code modifies the configuration data associated with the first program code by adjusting one or more levels of computing resources allocated for executing the first program code, the computing resources comprising one or more of CPU, memory, network, input/output (I/O), a maximum number of requests, or a maximum execution duration.

8. The system of claim 5, wherein the frontend computing system is further configured to:

determine a control policy associated with the first program code;

determine whether the control policy indicates that permission from an account holder associated with the first program code is needed before modifying the first program code; and

receive a confirmation from the account holder that the first program code may be modified in view of the detected trend, prior to modifying the at least one of the first program code or the configuration data associated with the first program code.

9. The system of claim 5, wherein the plurality of code execution requests are generated by an external service in response to an occurrence of a triggering event associated with the first program code, the external service comprising one of a storage service, a monitoring service, a logging service, a billing service, a data stream service, a messenger service, a database service, or a message queueing service.

10. A computer-implemented method comprising:

as implemented by one or more computing devices configured with specific executable instructions,

receiving a plurality of code execution requests associated with a first program code, each code execution request including one or more first input parameters to the first program code, the first program code associated with configuration data;

processing the plurality of code execution requests using the one or more first input parameters and the configuration data;

detecting a problem in executing one of the plurality of code execution requests associated with the first program code;

without successfully executing the first program code for the code execution request, routing the code execution request to a dynamic dead letter queue (DLQ) configured to store code execution requests;

causing at least one of the first program code or the configuration data associated with the first program code to be modified;

determining that the code execution request in the dynamic DLQ is to be re-processed;

re-processing the code execution request using the modified at least one of the first program code or the configuration data;

detecting a trend in the processing of the plurality of code execution requests;

selecting an action based on the detected trend; and

causing a second program code configured to perform the selected action to be executed on a pre-initialized virtual computing instance.

11. The computer-implemented method of claim 10, wherein the second program code modifies the first program code by updating one or more parameters associated with the first program code, the one or more parameters comprising at least one of version information, permission information, runtime information, or library information.

12. The computer-implemented method of claim 10, wherein the second program code modifies the configuration data associated with the first program code by adjusting one or more levels of computing resources allocated for executing the first program code, the computing resources comprising one or more of CPU, memory, network, input/output (I/O), a maximum number of requests, or a maximum execution duration.

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

determining a control policy associated with the first program code;

determining whether the control policy indicates that permission from an account holder associated with the first program code is needed before modifying the first program code; and

receiving a confirmation from the account holder that the first program code may be modified in view of the detected trend, prior to modifying the at least one of the first program code or the configuration data associated with the first program code.

14. Non-transitory physical computer storage comprising instructions that, when executed by one or more computing devices, configure the one or more computing devices to:

receive a plurality of code execution requests associated with a first program code, each code execution request including one or more first input parameters to the first program code, the first program code associated with configuration data;

process the plurality of code execution requests using the one or more first input parameters and the configuration data;

detect a problem in executing one of the plurality of code execution requests associated with the first program code;

without successfully executing the first program code for the code execution request, route the code execution request to a dynamic dead letter queue (DLQ) configured to store code execution requests;

cause at least one of the first program code or the configuration data associated with the first program code to be modified;

determine that the code execution request in the dynamic DLQ is to be re-processed;

re-process the code execution request using the modified at least one of the first program code or the configuration data;

detect a trend in the processing of the plurality of code execution requests;

select an action based on the detected trend; and

cause a second program code configured to perform the selected action to be executed on a pre-initialized virtual computing instance.

15. The non-transitory physical computer storage of claim 14, wherein the second program code modifies the first program code by updating one or more parameters associated with the first program code, the one or more parameters comprising at least one of version information, permission information, runtime information, or library information.

16. The non-transitory physical computer storage of claim 14, wherein the second program code modifies the configuration data associated with the first program code by adjusting one or more levels of computing resources allocated for executing the first program code, the computing resources comprising one or more of CPU, memory, network, input/output (I/O), a maximum number of requests, or a maximum execution duration.

17. The non-transitory physical computer storage of claim 14, wherein the instructions further configure the one or more computing devices to:

determine a control policy associated with the first program code;

determine whether the control policy indicates that permission from an account holder associated with the first program code is needed before modifying the first program code; and

receive a confirmation from the account holder that the first program code may be modified in view of the detected trend, prior to modifying the at least one of the first program code or the configuration data associated with the first program code.