Abstract: Techniques are described for performing different types of restart operations for a file storage replication between a source file system and a target file system in different cloud infrastructure regions. In certain embodiments, the disclosed techniques perform a restart operation to terminate a current cross-region replication by synchronizing resource cleanup operations in the source file system and the target file system, respectively. In other embodiments, disclosed techniques perform a restart operation to allow a customer to reuse the source file system by identifying a restartable base snapshot in the source file system without dependency on the target file system.

Abstract: Novel techniques for end-to-end file storage replication and security between file systems in different cloud infrastructure regions are disclosed herein. In one embodiment, a file storage service generates deltas between snapshots in a source file system, and transfers the deltas and associated data through a high-throughput object storage to recreate a new snapshot in a target file system located in a different region during disaster recovery. The file storage service utilizes novel techniques to achieve scalable, reliable, and restartable end-to-end replication. Novel techniques are also described to ensure a secure transfer of information and consistency during the end-to-end replication.

Abstract: Techniques are described for ensuring end-to-end fair-share resource utilization during cross-region replication. In certain embodiments, a fair-share architecture is used for communication among pipeline stages performing a cross-region replication between different cloud infrastructure regions. Cross-region replication-related jobs are distributed evenly from a pipeline stage into a temporary buffer in the fair-share architecture, and then further distributed evenly form the fair-share architecture to parallel running threads of next pipeline stage for execute. Techniques for static and dynamic resource allocations are also disclosed.

Abstract: Techniques are described for performing different types of restart operations for a file storage replication between a source file system and a target file system in different cloud infrastructure regions. In certain embodiments, the disclosed techniques perform a restart operation to terminate a current cross-region replication by synchronizing resource cleanup operations in the source file system and the target file system, respectively. In other embodiments, disclosed techniques perform a restart operation to allow a customer to reuse the source file system by identifying a restartable base snapshot in the source file system without dependency on the target file system.

Abstract: Techniques are described for partitioning B-tree keys of file systems into key ranges for parallel processing in delta generation during file storage replications between file systems in different cloud infrastructure regions. In certain embodiments, a delta generation processing for cross-region replication may utilize a key-range splitting mechanism involving a recursive algorithm that partitions B-tree keys of a source file system into roughly equal-size key ranges. All the partitioned key ranges may be processed in parallel and concurrently by different processing threads, one thread per key range, to improve the performance of the delta generation and achieve scalability.

Abstract: Systems and methods for packet payload mapping for robust transmission of data are described. For example, methods may include receiving, using a network interface, packets that each respectively include a primary frame and one or more preceding frames from the sequence of frames of data that are separated from the primary frame in the sequence of frames by a respective multiple of a stride parameter; storing the frames of the packets in a buffer with entries that each hold the primary frame and the one or more preceding frames of a packet; reading a first frame from the buffer as the primary frame from one of the entries; determining that a packet with a primary frame that is a next frame in the sequence has been lost; and, responsive to the determination, reading the next frame from the buffer as a preceding frame from one of the entries.

Abstract: Techniques are described for efficient replication and maintaining snapshot data consistency during file storage replication between file systems in different cloud infrastructure regions. In certain embodiments, provenance IDs are used to efficiently identify a starting point (e.g., a base snapshot) for a cross-region replication process, conserve cloud resources while reducing network and IO traffic.

Abstract: Techniques are described for efficient replication and maintaining snapshot data consistency during file storage replication between file systems in different cloud infrastructure regions. In certain embodiments, snapshot creation and deletion requests that occur during cross-region replications may be temporarily withheld until appropriate times to execute such requests safely, depending on the timing relationship between such requests and cross-region replication cycles.

Abstract: Novel techniques are described for performing hierarchical key management involving an end-to-end file storage replication between different cloud infrastructure regions. The novel hierarchical key management comprises three different keys, a first security key for the source region, a session key, valid only for a session, for the transfer of data between two different regions, and a second security key for the target region. Novel techniques are also described for using different file keys for different files of a file system in each region.

Abstract: A method and a memory device are provided. A controller of the memory device retrieves a first command from a host memory. The controller stores a first identifier of the first command in a set-aside buffer of the memory device. The controller retrieves a second command from the host memory.

Abstract: Methods and systems for processing a command from a host to a storage device are disclosed. A first controller of the storage device receives a first command from the host via a first queue. The first queue is exposed to the host. The storage device generates a second command based on the first command, and submits the second command to a second controller of the storage device via a second queue. The second command and the second queue are unexposed to the host. The second controller obtains and processes the second command from the second queue, where the processing of the second command is for accessing non-volatile storage media of the storage device. Based on a status of the second command, the first controller transmits a signal to the host for indicating completion of the first command.

Abstract: System and method for differentiable networks trainable to learn an optimized query of a 3D model database used for object recognition includes training a first differentiable network configured as a differentiable renderer by generating 2D images from 3D models of a first object of a dissimilar second object while optimizing rendering parameters for producing 2D images by gradient descent of a first triple loss function. Visual variation among the images is maximized. A second differentiable network configured as a convolutional neural network defined by a regression function is trained by generating searchable feature vectors of the 2D images. The feature vectors are determined using optimized neural network parameters determined by gradient descent of a second triple loss function to achieve high correlation to an input image of the first object and low correlation to images of the second object.

Abstract: Systems and methods for packet payload mapping for robust transmission of data are described. For example, methods may include receiving, using a network interface, packets that each respectively include a primary frame and one or more preceding frames from the sequence of frames of data that are separated from the primary frame in the sequence of frames by a respective multiple of a stride parameter; storing the frames of the packets in a buffer with entries that each hold the primary frame and the one or more preceding frames of a packet; reading a first frame from the buffer as the primary frame from one of the entries; determining that a packet with a primary frame that is a next frame in the sequence has been lost; and, responsive to the determination, reading the next frame from the buffer as a preceding frame from one of the entries.

Abstract: Systems and methods for playing synchronized audio streams for live broadcasts are described. For example, supplemental audio streams (e.g., extra commentary or sound effects) may be streamed over a packet switched network with a copy (e.g., a reduced bandwidth copy) of a main audio signal from a television broadcast. A user device may synchronize the streamed audio with the audio from a television recorded with a microphone, and play a supplemental audio signal synchronously with the audio from the television. For example, methods may include capturing a first audio signal using a microphone; receiving a second audio signal and a third audio signal using a network interface; determining an estimated delay of the first audio signal relative to the second audio signal; and playing, using a speaker, the third audio signal with an added delay determined based on the estimated delay.

Abstract: A method for depositing a protective coating onto a substrate, wherein the protective coating comprises (i) a moisture-barrier layer which is in contact with the substrate and which comprises a first sub-layer, optionally one or more intermediate sub-layers, and a final sub-layer, (ii) a mechanical-protective layer which is inorganic, and (iii) a gradient layer interposing the moisture-barrier layer and the mechanical-protective layer.

Abstract: Systems and methods for processing copy commands are disclosed. A first controller of the storage device receives a copy command from a host via a first queue. The storage device generates, based on the copy command, a read command and a write command, and submits the read and write commands to a second controller of the storage device via a second queue. The second controller retrieves and processes the read and write commands from the second queue. The storage device reads, based on the processing of the read command, data stored in a first location of a storage media associated with a source address, and writes the data to a second location of the storage media associated with a destination address. The first controller transmits a signal to the host for indicating completion of the copy command.

Abstract: Anomalies in multiple different scenes or images can be detected and localized in a single training flow of a neural network. In various examples, incremental learning can be applied to a given system or network, such that the system or network can learn the distribution of new scenes in a single training flow. Thus, in some cases, when an anomalous image from a new scene is given as input to the network, the network can detect and localize the anomaly.

Abstract: The disclosure includes a system and method for receiving, using one or more processors, first image data representing a first image of a document; obtaining, using the one or more processors, a first set of corner coordinates representing the corners of the document in the first image; generating, using the one or more processors, a first polygon based on the first set of corner coordinates; performing, using the one or more processors, one or more corner-based validity checks based on the first set of corner coordinates; and performing, using the one or more processors, an auto-capture.

Abstract: A computerized method of delivering data to one or more client computing devices includes receiving, by a server computing device, a data stream starting at a first time; processing, by the server computing device, the data stream, thereby creating a processed data stream; transmitting, by the server computing device, via a wireless network in electronic communication with the server computing device, the processed data stream to the one or more client computing devices; and interpreting, by an application installed on the one or more client computing devices, the processed data stream, thereby recovering the data stream for use by the one or more client computing devices. A latency between the first time and the second time is less than 100 milliseconds.

Abstract: A computer-implemented method can comprise accessing a trained learning machine, evaluating, using the machine learning model, the transcript to output a first sentiment score related to the first party in the unique domain, accessing digital engagement data representing engagement of the first party with digital assets associated with the second party, evaluating the one or more sentiment score values and the digital engagement data to output a value indicative of a likelihood of the first party to take a particular action, and determining whether the value is above a threshold, and if so, automatically sending a notification to a computer device associated with the second party.