Amazon Technologies, Inc.

Matthew M Luker

David Allen Markley

Christopher Vincent Kaukl

Chris Predeek

Alfred Yong-Hock Tan

A representation of a workflow comprising a plurality of tasks is obtained. An execution of an instance of the workflow is initiated. The execution comprises selecting, with respect to a particular task of the workflow, a particular execution resource option from a set comprising at least a first execution resource option and a second resource execution option. A result of the execution is stored.

Dynamic execution resource selection for customized workflow tasks

Techniques are disclosed herein for selection of an electronic system to perform a transaction requested from within software. Using technologies described herein, software can utilize different electronic systems to perform a transaction. For example, software may allow a user to acquire processing by a remote system, such as obtaining additional content or services from more than one electronic system. Instead of software being tied to a particular electronic system (e.g., the electronic system from which the software was acquired), the software can identify and provide options of different electronic systems for performing a transaction. A software provider can use a software developer kit and/or application programming interfaces provided by a service provider to incorporate functionality to identify and access the different electronic systems that are available for selection.

Dynamic identification and selection of application programming interface

A translation service is used to modify service requests by translating client-provided external identifiers for objects into canonical identifiers for the objects (e.g., an identifier used by the service to uniquely identify the one or more objects). The translation service first receives specification data for a particular service. The specification data includes information for translating external identifiers for one or more objects to canonical identifiers for the one or more objects. The translation service instantiates a translation layer for the particular service, based on the received specification data. The translation layer receives a request directed to the particular service, translates the external object identifiers to canonical object identifiers, adds the canonical object identifiers to the request, and forwards the request to the particular service. The translation layer may also remove one or more of the external object identifiers from the request before sending to the particular service (for data isolation).

Transparent support and translation for client-provided identifiers in service requests

Methods and systems for receiving shouted-out user responses to broadcast entertainment content, and for determining the responsiveness of those responses in relation to the broadcast content. In particular, entertainment broadcasts can be accompanied by mark-up data that represents various events within a given broadcast, which can be compared to the shouted-out responses to determine their accuracy. For example, if a game show was broadcast and an individual started shouting out answers during the broadcast, embodiments disclosed herein could utilize a voice-controlled electronic device that captures the shouted-out answers and passes them on to a language processing system that determines whether they are correct by comparing the answers to the mark-up data. The voice-controlled electronic device can also 'listen' to background sounds to capture the broadcast of the entertainment content, and send that content to the language processing system, which can use that captured data to synchronize the actual broadcast with the analysis of the shouted-out answers to provide individuals with an immersive entertainment experience.

Growing community of inventors

Bainbridge Island, WA, United States of America

Matthew M Luker