Amazon Technologies, Inc.

Ramyanshu Datta

Kevin Christopher Miller

Timothy Lawrence Harris

Andrea Olgiati

Saurabh Gupta

Lakshmi Naarayanan Ramakrishnan

Wei You

Kohen Berith Chia

Kunal Mehrotra

Arvind Sowmyan

Ishaaq Chandy

Robert Devers Wilson

Haripriya Devnath

Arun Babu Nagarajan

Chandan Talukdar

Zhihan Li

A post-task-completion retention period for which a computing resource is to be retained, without de-activating the resource, on behalf of a set of requesters of machine learning tasks is determined at a machine learning service. A first task, identified at the service prior to expiration of the retention period at a first computing resource at which a second task has completed, is initiated at the first computing resource. In response to obtaining an indication of a third task and determining that a threshold criterion associated with the retention period satisfies a criterion, the third task is initiated at an additional computing resource. The additional computing resource is de-activated after the third task completes, without waiting for the retention period to expire.

Client-configurable retention periods for machine learning service-managed resources

Parameters of a pool of computing resources to be utilized for machine learning tasks from a set of entities are stored, including a category of the computing resources, and a post-task-completion retention period during which, after completion of a task, at least a portion of data stored at the resource is not to be deleted. A compute instance of the pool is assigned to a task requested from the set of entities after determining that one or more configuration settings of the instance satisfy a preference indicated in the request for the task, and that the retention period of the instance relative to a completion of an earlier task on the instance has not expired. A result of the task is stored.

Resource management techniques to reduce startup overhead for machine learning tasks

Systems and methods are described for modifying input and output (I/O) to an object storage service by implementing one or more owner-specified functions to I/O requests. A function can implement a data manipulation, such as filtering out sensitive data before reading or writing the data. The functions can be applied prior to implementing a request method (e.g., GET or PUT) specified within the I/O request, such that the data to which the method is applied may not match the object specified within the request. For example, a user may request to obtain (e.g., GET) a data set. The data set may be passed to a function that filters sensitive data to the data set, and the GET request method may then be applied to the output of the function. In this manner, owners of objects on an object storage service are provided with greater control of objects stored or retrieved from the service.

On-demand execution of object transformation code in output path of object storage service

Systems and methods are described for modifying input and output (I/O) to an object storage service by implementing any combination of and any number of owner-specified functions and native functions. A function can implement a data manipulation. The functions can be applied prior to implementing a request method (e.g., GET, PUT, LIST, etc.) specified within the I/O request, such that the data to which the method is applied may not match the object specified within the request. For example, a user may request to obtain a data set. The data set may be passed to a native function that filters sensitive data to the data set, the output of the native function may be passed to an owner-specified function that redacts data from the filtered data set, and the request method may then be applied to the output of the owner-specified function.

Sequential execution of user-submitted code and native functions

Systems and methods are described for modifying input and output (I/O) to an object storage service by implementing one or more owner-specified functions to I/O requests. Such functions can include data access control functions, data manipulation functions, and the like. The owner of an object collection maintained by the object storage service can specify code execution environment rules that can give privileges to the execution of such functions such as by allowing the functions to access external services or the requesting user's private resources. In this manner, owners of the object collection are provided with greater control over how the object collection is accessed.

Code execution environment customization system for object storage service

Systems and methods are described for modifying input and output (I/O) to an object storage service by implementing one or more owner-specified functions to I/O requests. Different data manipulation functions can be placed in different I/O paths depending on the request method or user access level. For example, a user having full access may be returned the unaltered version of the object, whereas a user having modified or reduced access may be returned a modified or redacted version of the object. In this manner, owners of the object collection are provided with greater control over how the object collection is accessed.

User-specific data manipulation system for object storage service based on user-submitted code

Systems and methods are described for modifying input and output (I/O) to an object storage service by implementing one or more owner-specified functions to I/O requests. A function can implement a data manipulation, such as filtering out sensitive data before reading or writing the data. The functions can be applied prior to implementing a request method (e.g., GET or PUT) specified within the I/O request, such that the data to which the method is applied my not match the object specified within the request. For example, a user may request to obtain (e.g., GET) a data set. The data set may be passed to a function that filters sensitive data to the data set, and the GET request method may then be applied to the output of the function. In this manner, owners of objects on an object storage service are provided with greater control of objects stored or retrieved from the service.

On-demand code obfuscation of data in input path of object storage service

Inserting owner-specified data processing pipelines into input/output path of object storage service

On-demand execution of object combination code in output path of object storage service

Execution of owner-specified code during input/output path to object storage service

Inserting executions of owner-specified code into input/output path of object storage service

Systems and methods are described for modifying input and output (I/O) to an object storage service by implementing one or more owner-specified functions to I/O requests. A function can implement data access control, such as controlling which users are provided access to which portions of an object collection maintained by the object storage service. For example, data access control functions can be applied prior to implementing a request method (e.g., GET or PUT) specified within the I/O request, and may grant or deny access based on a variety of factors such as user identity, time window, prior access, keywords, geographical region, etc. In this manner, owners of the object collection are provided with greater control over how the object collection is accessed.

Growing community of inventors

Seattle, WA, United States of America

Ramyanshu Datta