Abstract: A transcoding service is described that is capable of transcoding or otherwise processing content, such as video, audio or multimedia content, by utilizing one or more pipelines. A pipeline can enable a user to submit transcoding jobs (or other processing jobs) into an available pipeline, where a transcoding service (or other such service) assigns one or more computing resources to process the jobs received to each pipeline. The transcoding service and the pipelines can be provided by at least one service provider (e.g., a cloud computing provider) or other such entity to a plurality of customers. A service provider can also provide the computing resources (e.g., servers, virtual machines, etc.) used to process the transcoding jobs from the pipelines.

Abstract: In response to a first programmatic request, metadata indicating that a first isolated read channel of a real-time category has been associated with a first target stream is stored at a stream management service. In response to another request, metadata indicating that a second isolated read channel of a non-real-time category has been associated with a second target stream is stored. In response to a read request indicating the first channel or the second channel, one or more data records of the corresponding target streams are provided.

Abstract: The application describes an authentication process that incorporates voice commands with an HTTP interface module to perform a multifactor authentication (MFA) process. For example, a first computer system may initiate the MFA process by sending, with a HTTP interface module maintained by the first computer system, a request to initiate the MFA process to the second computer system. The MFA process may also transmit an one-time password (OTP) to a first user device. The first computer system may receive an audible recitation of the OTP from a second user device. The OTP may be parsed and used to generate an HTTP request. The HTTP interface module may send the HTTP request to the second computer system. When the second computer system authenticates the user based at least in part on the non-audible file, the first computer system may receive confirmation of authentication of the user to initiate the transaction.

Abstract: A system may route incoming data to different data shards using a scalable system architecture. The system may include a data ingestion service, such as a naming service that may input data for you by the system. This data may be routed to specific applications, which may be associated with different shards of the data. The system may decouple data ingestion components from application components to enable adding additional capacity at either end of the system, or anywhere between, to accommodate increases in demand and/or to increase throughput. For example, the system may be modified to add additional data ingestion components without modifying the number of applications components to increase throughput of data ingestion. Similarly, the system may be modified to add application components without modifying the number of data ingestion components to increase throughput of data analysis, for example.

Abstract: A secret is stored in a computing device. The device generates a value determined based at least in part on a substantially random process. As a result of the value satisfying a condition, the device causes the secret to be unusable to perform cryptographic operations such that the device is unable to cause the secret to be restored. The secret may be programmatically unexportable from the device.

Abstract: Commitments against various resources can be dynamically adjusted for customers in a shared-resource environment. A customer can provision a data volume with a committed rate of Input/Output Operations Per Second (IOPS) and pay only for that commitment (plus any overage), for example, as well as the amount of storage requested. The customer can subsequently adjust the committed rate of IOPS by submitting an appropriate request, or the rate can be adjusted automatically based on any of a number of criteria. Data volumes for the customer can be migrated, split, or combined in order to provide the adjusted rate. The interaction of the customer with the data volume does not need to change, independent of adjustments in rate or changes in the data volume, other than the rate at which requests are processed.

Abstract: Techniques are described for facilitating use of invocable services by applications in a configurable manner. In at least some situations, the invocable services are Web services or other network-accessible services that are made available by providers of the services for use by others in exchange for fees defined by the service providers. The described techniques facilitate use of such invocable services by applications in a manner configured by the developers of the applications, including to allow the application developers to configure pricing terms that will be used to determine fees that end users of the applications will be charged for use of the invocable services via the applications. In some situations, the configured pricing terms for a service specify fees for end users that differ in one or more ways from the defined fees charged by the provider of that service.

Abstract: Systems and methods are described to enable management of redundant route announcements in an access point including multiple packet processors. Route controllers are described that can generate routing information distributing incoming packets to the access point among the packet processors. The route controllers can operate redundantly, such that a failure of a single controller does not cause a complete failure of the access point. To avoid different announcements by different route controllers (particularly under partial failure scenarios), the route controllers utilize a strongly consistent data store to store routing information. So long as a record within the data store contains valid information, it is considered authoritative and routing information from the record is announced by all route controllers. If the information is invalid, the route controllers attempt to overwrite the information with new, higher priority routing information.

Abstract: An automatic speech recognition (ASR) system detects an endpoint of an utterance based on a domain of the utterance. The ASR system processes a first portion of the utterance to determine the domain and then determines an endpoint of the remainder of the utterance depending on the domain.

Abstract: Methods and apparatus for centralized resource utilization visualization in distributed systems are disclosed. Metrics are collected from a plurality of sources in a distributed system, including network traffic metrics collected from nodes implementing a network-accessible service. A topology representing at least one of (a) a relationship between respective client accounts to which a first and second node are assigned, or (b) one or more network links between the first and second nodes is generated. Respective networking performance indicators of the first node and the second node are provided for inclusion in a resource heat map corresponding to the topology.

Abstract: Techniques for optimizing a system to improve an overall user satisfaction in a speech controlled system are described. A user speaks an utterance and the system compares an expected sum of user satisfaction values for each action to make a decision as to how best to process the utterance. As a result, the system may make a decision that decreases user satisfaction in the short term but increases user satisfaction in the long term. The system may estimate a user satisfaction value and associate the estimated user satisfaction value with a current dialog state. By tracking user satisfaction values over time, the system may train machine learning models to optimize the expected sum of user satisfaction values. This improves how the system selects an action or application to which to dispatch the dialog state and how a specific application selects an action or intent corresponding to the command.

Abstract: Technology is described for a checkpoint service for code function checkpoints associated with instances of code functions executed by a computing hub in a local area network. Subsequent to the execution of a code function being suspended by the computing hub, the computing hub may generate a code function checkpoint describing the execution of the code function. The computing hub may persist the code function checkpoint to storage using a checkpoint service. The computing hub may create the code function checkpoint in response to monitoring metrics associated with the computing hub, such as processor overutilization. The computing hub may resume the execution of the code function at a later time using the code function checkpoint retrieved from the checkpoint service.

Abstract: A device-management system performs audio processing, such as acoustic echo cancellation or beamforming, in a computing-resource allocation corresponding to a functionally limited device. The device-management system may be a locally-connected network device that is in communication with one or more user devices; the device-management system may also or instead be a remote device that communicates with the user devices using the locally-connected network device. The device-management system may receive audio data from one or more microphones of one or more user devices. To perform acoustic echo cancellation, the device-management system may receive and process time data corresponding to a time of output of audio by the user device.

Abstract: Systems and methods for device-state quality analysis are disclosed. Polling schedules may be generated based on which accessory devices have yet to be polled, which accessory devices are associated with less than favorable previous polling results, and/or which accessory devices are associated with a priority polling request. Polling requests may be generated using the schedules and the polling requests may be sent to systems associated with the accessory devices. Those systems may return current device-state data, which may be compared to stored device-state data to determine an accuracy value for the stored device-state data. Quality data may be generated using these accuracy values.

Abstract: Devices in an environment may be transitioned to a low power state based on calendar data or other types of user input. User input indicating a first device, a time period, and a set of conditions is received. When the time period occurs, device data from one or more devices in the environment is used to determine whether the set of conditions is met. If the conditions are met, an instruction from a device outside of the environment is provided to the first device to cause the first device to enter a low power state, during which the device is no longer accessible to other devices outside of the environment. After a lapse of the time period, a change in conditions, or in response to manual input provided to the first device, the first device returns to an operational state.

Abstract: It may be beneficial in the case of an undesired lapse of external power to provide a back-up power supply to protect electronic components in an electronic equipment rack. This can frequently be difficult and/or costly due to the necessary addition of electronic infrastructure, such as cabling and/or logic. One means of overcoming this obstacle is to utilize unused or reserved conductors in an already utilized management cable to convey a loss-of-power signal to the components in tandem with the existing cable signals.

Abstract: In one embodiment, a conveyor system has first and second conveyors. The first conveyor has inner and outer sides and a first conveyor surface between the inner and outer sides that is configured to carry payloads between first and second conveyor ends along a first direction. The second conveyor has inner and outer sides and a second conveyor surface between the inner and outer sides of the second conveyor that is configured to carry payloads between the first and second conveyor ends along the first direction. The first and second conveyors are coupled to one another such that the inner sides are spaced closer to one another than the outer sides and the first and second conveyors are angularly offset from one another. The conveyor system is configured to rotate between a first orientation and a second orientation about an axis that extends along the first direction.

Abstract: A system for providing a stateful virtual compute system is provided. The system may be configured to maintain a plurality of virtual machine instances. The system may be further configured to receive a request to execute a program code and select a virtual machine instance to execute the program code on the selected virtual machine instance. The system may further associate the selected virtual machine instance with shared resources and allow program codes executed in the selected virtual machine instance to access the shared resources.

Abstract: Embodiments of the present invention provide systems and methods for storing, accessing, and/or promoting previous versions of changed or overwritten content in a cloud computing or multitenant environment. According to various embodiments, the system may list all the versions, promote a version to be the current version of the node, delete a specific version of the node, and/or rename/move all the node versions. Metadata associated with file versioning may be stored in a dynamic table with concatenated identifiers as a hash key and a content version as another hash key. Accordingly, the method may include securely storing a copy of an object handle instead of copying the entire content at a new location, which reduces storage and resource requirements, as well as improve on security and performance by efficient lookups in the table for read, writes, updates, and deletes for the given node version.

Abstract: Implementations detailed herein include description of a computer-implemented method to migrate a machine learning model from one accelerator portion (such as a portion of a graphical processor unit (GPU)) to a different accelerator portion. In some instances, a state of the first accelerator portion is persisted, the second accelerator portion is configured, the first accelerator portion is then detached from a client application instance, and at least a portion of an inference request is performed using the loaded at least a portion of the machine learning model on the second accelerator portion that had been configured.

Abstract: Systems and methods for audio output control are disclosed. Audio may be output via a speaker of a communal device associated with a first portion of an environment. A user may provide a user utterance indicating an intent to add another device in a second portion of the environment to the audio-output session, and/or an intent to move the audio-output session from the first device to the second device, and/or an intent to remove a device from an audio-output session. Based on this determined intent, audio-session queues may be associated and dissociated from devices and device states may be altered to effectuate the intent of the user utterance.

Abstract: A service provider may provide a companion container instance associated with a mobile device in order to facilitate operation of the mobile device. The companion container instance and the mobile device may be associated in a database operated by the service provider. Furthermore, the companion container instance may execute various operations on behalf of the mobile diver based at least in part on a task definition indicating a software function to be executed by the companion container instance. The software function configured to execute the various operations on behalf of the mobile device.

Abstract: Static sorted index replication is described. A method may include receiving data to store in a memory tree of a replica in a partitioned distributed data store. The method may also include storing the received data in the respective memory tree of one of a plurality of replicas. The method may further include storing the received data sequentially in a static sorted index.

Abstract: Disclosed are various embodiments for software library versioning with intelligent caching based on signatures. In one embodiment, a library is requested in response to a first dependency in a first network resource. The library is then received together with a signature that identifies a version of the library. The library is stored in a local cache on a client computing device that is indexed by library signatures. It is determined that a second network resource has a second dependency on the library. The library is loaded from the local cache in response to the second dependency specifying the library having the same signature.

Abstract: A computing resource monitoring service receives a request to obtain data for various computing resources. The service obtains, from the various computing resources, one or more data log streams that include the requested data. The service utilizes the one or more data log streams to generate a probabilistic data structure that can be used to indicate that data log streams have been processed. If the one or more data log streams are not completely processed prior to the end of an allotted time period for processing of the request, the service generates a token that specifies partially processed data log streams and the probabilistic data structure. The token can be used to enable resumption of processing of the request.

Abstract: In response to determining that a secondary index is to be created for a particular table of a non-relational database service, a service component verifies that automated transmission of change records of the table to a log-structured journal has been configured. A given change record comprises a before-image and an after-image associated with a committed database write, and is assigned a unique sequence number. A change record propagator for the secondary index examines change records stored in the journal, and transmits selected change records to an index storage node. The index storage node builds the secondary index using the selected change records.

Abstract: A slave storage is provisioned using metadata of a master B-tree and updates to references (e.g., offsets) pertaining to data operations of the master B-tree. Master-slave pairs can be used to provide data redundancy, and a master copy can include the master B-tree with references to corresponding data. When provisioning a slave copy, the master sends a B-tree copy to the slave, which stores the slave B-tree copy, allocates the necessary space on local storage, and updates respective offsets of the slave B-tree copy to point to the local storage. Data from the master can then be transferred to the slave and stored according to a note and commit process that ensures operational sequence of the data. Operations received to the master during the process can be committed to the slave copy until the slave is consistent with the master and able to take over as master in the event of a failure.

Abstract: Systems and methods are directed to a computing device and methods for generating baselines of a data structure to enable selective reversion or editing of the data structure. A baseline may define a materialization of the data structure and may include pointers to immutable edit revisions associated with that materialization. The computing device may determine one or more change commands implemented to change a data structure from a first baseline to a second baseline. In response to receiving a request to revert at least part of the data structure from the second baseline to the first baseline, the computing device may identify the change command that was implemented to change that part of the data structure from the first baseline to the second line and may perform an inverse of the change command to revert that part of the data structure to a state corresponding to the first baseline.

Abstract: A data storage and retrieval system receives data points for a time series. The data storage and retrieval system stores the data points for a first portion of the time series using a first data format. Based at least in part on an analysis of queries performed on the first portion of the time series, the data storage and retrieval system determines to store a second portion of the time series using a second data format. The data storage and retrieval system stores subsequently received data points of the time series in the second data format.

Abstract: Systems, methods, and computer-readable media are disclosed for content delivery using persona segments for multiple users. In one embodiment, an example method may include receiving user interaction data from a user device for a session associated with a user account, wherein the user interaction data comprises indications of user interaction of interaction by a user account with a website. A set of keywords may be determined based on the product identifiers. A set of persona segments may be determined based on the keywords. Inter-persona distances between each persona segment of the set of persona segments may be determined. The inter-persona distances may be used to determine that multiple users are using a user account. A persona segment corresponding to a user may be selected. A content may be identified based on the persona segment.

Abstract: A database system may implement database system recovery using non-volatile system memory. An update to a data page of a database may be received. A version of the data page may be obtained in system memory. A new version of the data page may be determined according to the received update. The new version of the data page may be maintained in a non-volatile portion of system memory irrespective of a failure of the database. In at least some embodiments, the update may be performed without generating recovery log records indicating the update. Upon recovery from a database failure, data pages maintained in non-volatile system memory may be made available to service access requests.

Abstract: Methods and apparatus for lifecycle support for storage objects are disclosed. A storage medium stores program instructions that when executed on a processor implement a storage lifecycle manager of a multi-tenant storage service. The lifecycle manager determines a lifecycle policy to be implemented for a storage object. The policy indicates lifecycle transitions of the object, including a transition from a first service level to a second service level. The first service level and the second service level may each specify a maximum number of versions of the storage object. The lifecycle manager initiates modification operations corresponding to the storage object's data in accordance with the lifecycle policy, and verifies that the modification operations succeeded. The lifecycle manager provides, in response to a query, an indication of a current service level of the storage object.

Abstract: Systems and methods are described for providing auxiliary functions in an on-demand code execution system in a manner that enables efficient execution of code. A user may generate a task on the system by submitting code. The system may determine the auxiliary functions that the submitted code may require when executed on the system, and may provide these auxiliary functions by provisioning sidecar virtual machine instances that work in conjunction with the virtual machine instance executing the submitted code. The sidecars may provide auxiliary functions on a per-task, per-user, or per-request basis, and the lifecycles of the sidecars may be determined based on the lifecycles of the virtual machine instances that execute submitted code. Auxiliary functions may thus be provided only when needed, and may be provided securely by preventing a user from accessing the sidecars of other users.

Abstract: A wireless device includes a radio, a first directional antenna, a second directional antenna, an omnidirectional antenna, and a switch selectively coupled between the radio and the first directional antenna, the second directional antenna, and the omnidirectional antenna. A processor is coupled to the switch and to, for a frame: determine, based on an arbitration table, a destination medium access control address of a client wireless device and identifiers of antennas for transmitting the frame and receiving acknowledgement data; cause the switch to couple the radio to the first directional antenna; transmit the frame to the client wireless device via the first directional antenna, wherein the first client wireless device is located along a first direction with respect to the wireless device; cause the switch to couple the radio to the omnidirectional antenna; and receive an acknowledgment, corresponding to the frame, from the client wireless device via the omnidirectional antenna.

Abstract: A selective data restoration technique is disclosed. A request to recover data in a dataset is received along with an identifier of a particular portion of the data to be recovered. The request may specify a time that indicates a desired recovery state (e.g., recover the values as they were at the time). Change data is determined for the specified time. For example, data stream filters generate recovery data by applying filters (e.g., keys) to a time window batch of the stream data. The stream data may correspond to changes to a data set, and the recovery data may correspond to only a portion of the dataset (e.g., a portion of a table). The system performs a recovery action, based on the recovery data, using the recovery data to automatically restore data in a data store, or providing the recovery date to the requesting entity, for example.

Abstract: A metadata catalog may store metadata about data sources. The metadata may define table, column, partition and/or data formats, as well as title information, textual descriptions, modification information, data schemas and access information, such as location and connection information. A metadata catalog service may provide a shared, persistent, central repository for the metadata. The metadata of the catalog may be published via multiple access mechanisms and programmatically obtained and processed by other programs via the metadata catalog service. Metadata consumers may utilize the metadata catalog service to perform various metadata tasks, such as metadata retrieval, creation, addition, modification and/or deletion. The metadata catalog service may be directly integrated into existing data analytic tools as a persistent and durable metadata store allowing users to utilize the broad set of data processing tools.

Abstract: Techniques for providing customer feedback related to virtual machine instance maintenance events are described herein. A customer is provided with an event message specifying a virtual machine maintenance operation and a proposed time to perform that virtual machine maintenance operation. A response that specifies the time that the customer wishes to perform the virtual machine maintenance operation is received and the virtual machine maintenance operation is scheduled based at least in part on that specified time.

Abstract: Described are techniques for determining a data policy suitable for association with a data object based on the data access pattern for the data object. Correspondence between the data access pattern of the data object and pattern data, indicative of data access patterns stored in association with data policies, may be determined. Based on the correspondence between the data access pattern of the data object and a particular data access pattern of the pattern data, the data policy associated with the particular data access pattern may be suitable for use with the data object. A set of suitable data policies may be refined based on the content or metadata associated with the data object and the code or deployment status of services that access the data object. Once the access pattern for a data object is known, subsequent interactions with the data object may be analyzed to identify anomalous traffic.

Abstract: Techniques for curating audio and IR commands using machine learning may be provided. For example, the system can receive an audio stream that includes a plurality of audio segments provided by a particular user (e.g., as determined through acoustic fingerprinting) and the system can store the audio stream and/or segments in the user's profile. The system can also store a command associated with the audio segment(s). When a portion of the audio is provided in conjunction with a same command over a threshold number of repetitions, the next time that that audio segment is received, the system may provide the command that corresponds with that audio segment to an output device to cause an operation of the output device.

Abstract: Described are systems and techniques configured to present information to a user of a materials handling facility that stores items in inventory locations. Some or all of the information may be designated as confidential or have restrictions on presentation to reduce or eliminate dissemination of the confidential information to unauthorized users. A device such as a display carried by the user, mounted on a tote or shelf, and so forth may present the confidential information on a user interface. Presentation of the confidential information, acceptance of user inputs to the user interface, or both may be based on identification of an authorized user, a relative position between the user and the display, or other factors.

Abstract: A digital work may be annotated using an eBook reader device. Upon receiving an annotation relating to a specific portion of the digital work, an invariant location reference identifier corresponding to the specified portion of the digital work may be appended to the annotation. The annotation may then be stored in association with the digital work for later reference. In some instances, an annotation may be presented on an eBook reader device upon receipt of a valid authorization credential granting access to the annotation.

Abstract: Described are systems, methods, and apparatus for detecting objects within a distance of an aerial vehicle, and developing a three-dimensional model or representation of those objects. Rather than attempting to use stereo imagery to determine distances and/or depth of objects, the described implementations utilize range-gating, or time-gating, and the known position of the aerial vehicle to develop a three-dimensional representation of objects. For example, when the aerial vehicle is at a first position it may use range-gating to detect an object at a defined distance from the vehicle. The aerial vehicle may then alter its position and use range-gating to detect an object that is the defined distance from the vehicle at the new position. This may be done at several different positions and the resulting information and aerial vehicle position information combined to form a three-dimensional representation of those objects.

Abstract: A cloud wallet system may receive an indication from a third party system that a user has checked-in at a merchant. The cloud wallet system then determines whether the merchant has an account with the cloud wallet system so that a point-of-sale (POS) purchase can be processed via the user check-in. If the merchant has an account with the cloud wallet system, a user profile corresponding to the user is sent to the merchant and a point-of-sale transaction based on a third party user check-in may be processed by the merchant through the cloud wallet system.

Abstract: Described herein are a system and methods for providing action recommendations to a user that are likely to result in performance of a high-value action. In some embodiments, the user is compared to one or more other users in order to identify high-value actions for that user. Once at least one high-value action has been identified, a sequence of actions may be generated to include that high-value action using prediction model data that includes probability information. The sequence of actions is then assessed to determine a gateway action within the sequence of actions that is likely to be performed by the user and has a high likelihood of resulting in subsequent performance of the high-value action. The gateway action may then be presented to the user.

Abstract: Methods, systems, and computer-readable media for management of virtual desktop instance pools are disclosed. A plurality of virtual desktop instances are provisioned in a pool for a client organization. The number of virtual desktop instances does not exceed a number of virtual desktop slots for the client organization. To a first client device associated with a first user, access is provided to a particular virtual desktop instance based (at least in part) on a determination that a current number of connected virtual desktop instances is less than the number. To a second client device associated with a second user, access is denied to the plurality of virtual desktop instances based (at least in part) on a determination that a current number of connected virtual desktop instances meets the number.

Abstract: Disclosed are various embodiments for facilitating item bundles. An item bundle includes more than one item from an item universe that can be bundled together. For example, in an electronic marketplace, two items can be linked together as a bundle that is sold. An order document can be created that includes a reference to the item bundle that is associated with an entry for each of the items in the bundle.

Abstract: Techniques for identifying an automatic replenishment device (ARD) to item pairing based on an interaction by a user with a component of the ARD are disclosed herein. In embodiments, information from an ARD may be received that corresponds to an identify request based at least in part on an interaction with a component of the ARD. The information may include a device identifier for the ARD and a registration token. ARD data that identifies the ARD and the item associated with the ARD may be determined based at least in part on the received information. In accordance with at least one embodiment, a notification may be generated based at least in part on the ARD data where the notification is configured to present the item associated with the ARD. The notification may be transmitted to a user device associated with the ARD.

Abstract: Supplemental content is selected or generated based at least in part upon colors of primary content with which the supplemental content is to be displayed. Color data is determined for primary content and that color data is used to select supplemental content that includes complementary or similar colors. Past performance data can be analyzed in order to determine which colors are most effective for a type of opportunity. When an opportunity arises to provide supplemental content, the effective colors can be recommended or, in some cases, automatically applied such that the supplemental content will include colors that improve the overall likelihood of performance of the supplemental content. If a designer is generating supplemental content, a tool can use such performance data to recommend colors based on the type of supplemental content or other such targeting criteria.