Microsoft Technology Licensing, LLC

Vmware, Inc.

Linkedin Corporation

Red Hat, Inc.

Salesforce Com, Inc.

Vettery, Inc.

Speechace LLC

Abhishek Gupta

Anmol Bhasin

Richard Paul Spillane

Kapil Chowksey

Jorge Guerra Delgado

Christian Posse

Srinath Premachandran

Sandeep Rangaswamy

Wenguang Wang

Robert T Johnson

Rob T Johnson

Vijay K Dialani

Shivakumar Thangapandi

Vikrant Arora

Xianren Wu

Chun Ho Cheung

Ran Kalach

Shakti Dhirendraji Sinha

Jian H Li

Sachin Rekhi

Ye Xu

Angela Yoonjeong Yang

Satya Kanduri

Yan Yan

Yurong Shi

Annabel Fang Liu

Jacob Kessler

Monica Rogati

Ittai Abraham

James Robert Benton

Viet Thuc Ha

Sahin Cem Geyik

Jiong Wang

Ramesh Dommeti

Junghoon (Andrew) Ahn

Alexis Pribula

Zang Li

Jin Zhang

Paul Adrian Oltean

Dhanurjay A S Patil

Peter Nicholas Skomoroch

Matthew Terence Hayes

Radhika Vullikanti

Dmytro Andriyovich Ivchenko

Elizabeth Ethel Burstein

Kashif Hasan

Deepak Narula

Parker R Barrile

Jin Li

Christos Karamanolis

Sudipta Sengupta

Christopher Brian Quirk

Maxime Austruy

Pascal Renauld

Daniel McPherson

Raman Chandrasekar

Kun Liu

Jaewon Yang

Matthew R Richardson

Ping Xie

Sarthak D Shah

Mohit Kumar Garg

Abhishek Tiwari

Shan Zhou

Ronakkumar N Desai

Anish Ramdas Nair

Ahmed El-Shimi

Chenguang Zhu

Cagri Ozcaglar

Ashok Kumar Nandoori

Jordan Thomas Liggitt

Ning Wei

Rahim Daya

Joerg-Thomas Pfenning

Daniel Hefenbrock

Jui-Ting Huang

Junghoon Andrew Ahn

Christopher John Champness Burges

Amit Kumar Nanda

Zal Yaz Bilimoria

Gunjan Bansal

Zian Yu

Jordan Anthony Saints

Timothy Patrick Jordt

Gregory Alan Walloch

Zachary Tyler Piepmeyer

Chun Ho (Ian) Cheung

Ahmet Bugdayci

Hao Ma

Akshat Kale

Amol Wate

Nilambari Narayan Deshpande

Simran Rajkumar Nagrani

Bhavish Agarwal

Arun Venkatachalam

Siegfried Joseph Bilstein

Viraj Turakhia

Automatically selecting an encrypted network connection (such as a VPN tunnel) to use when communicating through another encrypted connection node with which there are multiple encrypted network connections with the computing system. The selection is based predicted latency of that encrypted network connection. The prediction is based on round-trip times of communications over the respective encrypted network connection. This is quite difficult since the encrypted network connection itself is not the entire path of the outgoing message and corresponding incoming message.

Automatic selection of encrypted network connection based on predicted latency

The techniques described herein increase the throughput of a single VPN connection by creating multiple outbound and/or inbound Security Associations (SAs). For instance, two or more different SAs can encrypt outbound data packets to be sent over the VPN connection to a remote device. Moreover, two or more different SAs can decrypt inbound data packets received over the VPN connection from the remote device. Each of the SAs can be bound to a different processing core via the use of a Security Parameter Index (SPI) identifier. Consequently, inbound data packets communicated over a single VPN connection from a remote device to a physical host in a VPN gateway can be distributed amongst multiple processing cores for decryption purposes. Further, outbound data packets to be communicated over the single VPN connection from the physical host to the remote device can be distributed amongst multiple processing cores for encryption purposes.

Throughput for a single VPN connection using multiple processing cores

A virtual private network (VPN) gateway selects a VPN connection, from multiple VPN connections established between a network VPN gateway and a remote VPN gateway, based on performance factors such as throughput. A system may measure throughput in megabytes per second (Mbps). More specifically, a VPN gateway (e.g., a remote VPN gateway or a network VPN gateway) can configure a routing preference that selects a VPN connection that is more performant based on a cryptographic algorithm that is used for the VPN connection. The VPN gateway can update the routing preference to select an alternative VPN connection when the performance of the VPN connection suffers.

Selecting a VPN connection using negotiated cryptographic algorithms to improve throughput

The presently disclosed subject matter includes an apparatus with a processor and a memory storing code which, when executed by the processor, causes the processor to receive a data profile associated with a candidate resource, the data profile includes a set of attributes of the candidate resource which are relevant for assessing the candidate resource's suitability to satisfy a particular resource demand. The apparatus extracts an n-dimensional feature vector from the received data profile, the n-dimensional feature vector capturing aspects of the candidate resource's attributes and process said n-dimensional feature vector with a first ensemble machine learning model to generate a first suitability factor. Likewise, the apparatus process said n-dimensional feature vector with a second ensemble machine learning model to generate a second suitability factor. The apparatus determines whether to allocate the candidate resource to the particular resource demand using said first and second suitability factors.

System for the computer matching of targets using machine learning

Exemplary methods, apparatuses, and systems include a file system process reading a first node in a tree data structure from a first memory. The first node includes a first approximate membership query data structure ('AMQ'), a first plurality of child pointers, a first plurality of pivot values, and a first buffer. The file system process determines that the first plurality of child pointers exceeds a maximum size. Using a pivot value in the first plurality of pivot values, the file system process splits the first node into a second node and a third node. The file system process uses the pivot value to split the first buffer into a second buffer and a third buffer. Using the pivot value and the first AMQ, the file system process generates a second AMQ and a third AMQ.

Reduced bandwidth queries in B e-trees using approximate membership query data structures

The techniques described herein enable the establishment of two simultaneous virtual private network (VPN) connections for a VPN client operating on a remote computing device. The VPN client can establish first VPN connection with a first VPN server instance of a VPN gateway and a second VPN connection with a second VPN server instance of the VPN gateway. To establish two simultaneous VPN connections, the VPN client is configured to create and/or use two Transmission Control Protocol (TCP) sockets. In one example, a first VPN connection can be a primary VPN connection and a second VPN connection can be a dormant VPN connection configured as a backup in case of a service interruption with the first VPN connection. In another example, a data flow can be split across the first and second VPN connections, or alternate between using the first and second VPN connections, based on performance parameters.

Creating two simultaneous VPN connections for a VPN client

The techniques described herein enable the establishment of two simultaneous virtual private network (VPN) connections between a VPN gateway and a VPN client. The system is configured to update a routing table advertised to network resources when a VPN server instance fails and/or is taken offline. When a first VPN server instance fails and/or is taken offline, the first VPN server instance releases a claim of ownership on its range of IP addresses. After this release occurs, the second VPN server instance is configured to claim ownership of the range of IP addresses that used to be owned by the first VPN server instance. This updated claim of ownership is captured in an updated routing table that can then be advertised to the network resources. Consequently, the network resources use this updated routing table to correctly determine which VPN server instance to send data intended for the VPN client.

Managing data traffic for two simultaneous VPN connections with a VPN client

Exemplary methods, apparatuses, and systems include a file system process inserting a first key/value pair and a second key/value pair into a first tree. The second key is a duplicate of the first key and the value of the second key/value pair is an operation changing the value. In response to a request for a range of key/value pairs, the process reads the second key/value pair and inserts it in a second tree. The process reads the first pair and determines, while inserting the first pair in the second tree, that the second key is a duplicate of the first key. The file system process determines an updated value of the first value by applying the operation in the second value to first value. The file system operation updates the second key/value pair in the second tree with the updated value and returns the requested range of key/value pairs.

Range lookup operations for B-trees using update messages

Embodiments described herein are related to a method of scanning a B-tree. For example, a method comprises receiving a scan request to scan a B-tree having a plurality of levels, each level comprising one or more nodes, wherein for each of one or more levels of the plurality of levels, nodes are grouped into groups, where nodes of any given group are stored across sequential disk blocks. The method further comprises generating a queue for each level of the B-tree. For each queue, the method further comprises loading into memory a next group of nodes based upon determining a storage location of a node of the next group of nodes.

Read iterator for pre-fetching nodes of a B-tree into memory

Techniques for enabling deduplication-aware load balancing in a distributed storage system are provided. In one set of embodiments, a node of the distributed storage system can receive an I/O (Input/Output) request pertaining to a data block of a storage object stored on a local storage component of the node. The node can further determine whether the I/O request requires insertion of a new entry into a deduplication hash table associated with the local storage component or deletion of an existing entry from the deduplication hash table. If the I/O request requires insertion of a new hash table entry, the node can add an identifier of the data block into a probabilistic data structure associated with the local storage component, where the probabilistic data structure is configured to maintain information regarding distinct data blocks that are likely present in the local storage component. Alternatively, if the I/O request requires deletion of an existing hash table entry, the node can remove the identifier of the data block from the probabilistic data structure.

Deduplication-aware load balancing in distributed storage systems

Embodiments of the present disclosure relate to techniques for performing a merge update for a database. In particular, certain embodiments of a method include generating a message comprising a first key and a first transaction associated with the first key, the first transaction indicating a transaction to perform other than for key-value pairs comprising the first key. The method further includes storing the message in a database. The method further includes merging the message with a first key-value pair stored in the database, the first-key value pair comprising the first key. The method further includes performing the first transaction based on merging the message with the first key-value pair.

Merge updates for key value stores

Exemplary methods, apparatuses, and systems include a file system process determining to a flush a node in a first tree. The first node includes a buffer structured as a second tree. The file system process generates an input/output instruction to load the buffer from a first memory to a second memory. The second tree is stored in two more non-contiguous locations in the first memory and the input/output operation includes a read operation corresponding to each of the two or more non-contiguous locations. The file system process causes the input/output instruction to be executed concurrently on the first memory.

Optimizing operations on b e-trees stored on solid state media

One or more implementations allow for systems, methods, and devices for teaching and/or assessment of one or more spoken language skills through analysis of one or more pronunciation characteristics of one or more individual language components of a teaching string audio sample data that corresponds to the user's speaking of the teaching string.

Teaching and assessment of spoken language skills through fine-grained evaluation

A B-tree associated with a file system on a storage volume includes a hierarchy of nodes. Each node includes a buffer portion to store key-value pairs as messages in the buffer. Each node can be characterized by having a maximum allowable size that is periodically updated at run time. The buffers in the nodes of the B-tree are therefore characterized by having a maximum allowed size that can vary over time.

Auto-tuned write-optimized key-value store

Growing community of inventors

San Francisco, CA, United States of America

Abhishek Gupta