Microsoft Technology Licensing, LLC

Yingnong Dang

Jian Wang

Dongmei Zhang

Qiang Wang

Liyong Chen

Xiaoxu Ma

Jiang Wu

Song Ge

Chunhui Zhang

Zhouchen Lin

Shi Han

Xiaohui Hou

Min Wang

Song Huang

Xinsheng Yang

Yue Li

Haidong Zhang

Cong Chen

Murali Mohan Chintalapati

Ken Hsieh

Sergio Paolantonio

Yantao Li

Xiang Cao

Rui Ding

Wenli Zhu

Si Qin

Pankaj K Singh

Adnan Azfar Mahmud

Justin Ding

Xiumin Lin

Sree Krishna Chaitanya Vadrevu

Ze Li

Zheng Chen

Fei Liu

Jue Wang

Hang Li

Qiang Fu

Jian Jun Wang

Bin Zhao

Xu Yang

Bing Jiang Sun

Kai Chen

Qian Wu

Xiao Bi Liang

Xiao Yang

Qingwei Lin

Tao Xie

Youjun Liu

Guowei Liu

Jim Sun

An Xu

Prashant Ratanchandani

Shuguang Ye

Yingjun Qiu

Yihua Xu

Hongyu Zhang

Scott Ruble

Qianchuan Zhao

Feng Liang

Shusen Liu

Haoyu Jiang

Youjiang Wu

Liyon Chen

Randolph Yao

Shulei Wang

Chenglin Zhong

Han Yin

Niko Schwarz

Chao Bian

Xiangpeng Zhao

Shuo-Hsien (Stephen) Hsiao

Techniques and systems for detecting leakage of computing resources in cloud computing architectures are described. In some implementations, first data may be obtained that indicates usage of a computing resource, such as non-volatile memory, volatile memory, processor cycles, or network resources, by a group of computing devices included in a cloud computing architecture. The first data may be used to determine reference data that may include a distribution of values of usage of the computing resource by the group of computing devices. Second data may also be collected that indicates usage of the computing resource by the group of computing devices during a subsequent time frame. The second data may be evaluated against the reference data to determine whether one or more conditions indicating a leak of the computing resource are satisfied.

Detection of computing resource leakage in cloud computing architectures

The present disclosure relates to systems, methods, and computer readable media for diagnosing and mitigating memory impact events, such as memory leaks, high memory usage, or other memory issues causing a host node from performing as expected on a cloud computing system. The systems described herein involve receiving locally generated memory usage data from a plurality of host nodes. The systems described herein may aggregate the memory usage data and determine a memory impact diagnosis based on a subset of the aggregated memory usage data. The systems described herein may further apply a mitigation model for mitigating the memory impact event. The systems described herein provide an end-to-end solution for diagnosing and mitigating a variety of memory issues using a dynamic and scalable system that reduces a negative impact of memory leaks and other memory issues on a cloud computing system.

Diagnosing and mitigating memory leak in computing nodes

A system for frequent pattern mining uses two layers of processing: a plurality of computing nodes, and a plurality of processors within each computing node. Within each computing node, the data set against which the frequent pattern mining is to be performed is stored in shared memory, accessible concurrently by each of the processors. The search space is partitioned among the computing nodes, and sub-partitioned among the processors of each computing node. If a processor completes its sub-partition, it requests another sub-partition. The partitioning and sub-partitioning may be performed dynamically, and adjusted in real time.

Frequent pattern mining

The techniques described herein provide tools that summarize a dataset by creating a final set of segments that, when visually presented via a histogram or other data presentation tool, show the distribution of at least a portion of the data. To create the final set of segments, the techniques described herein may collect or receive a dataset with distinct values, and divide the dataset into a number of segments that is less than or equal to a segment presentation threshold (e.g., ten segments). After creating the final set of segments, the techniques may configure and/or present data visualizations, such as histograms, for the created segments so that an observer is provided with a good viewing experience.

Data segmentation and visualization

Methods, systems and computer program products are described for obtaining deployment signals comprising information relating to deployments of software components to a plurality of computing devices, obtaining fault signals comprising information relating to faults encountered by the plurality of computing devices, and obtaining device type information that describes a device type of each of the plurality of computing devices. Based on the deployment signals, fault signals, and device type information, a correlation score for each combination of software component, fault, and device type is obtained. Based on the correlation scores, attribution level decisions, fault type level decisions and device type level decisions are made. Based on these decisions, an overall decision is rendered whether to proceed or not proceed with a future deployment of the software component.

Quality assessment and decision recommendation for continuous deployment of cloud infrastructure components

A system may include a node historical state data store having historical node state data, including a metric that represents a health status or an attribute of a node during a period of time prior to a node failure. A node failure prediction algorithm creation platform may generate a machine learning trained node failure prediction algorithm. An active node data store may contain information about computing nodes in a cloud computing environment, including, for each node, a metric that represents a health status or an attribute of that node over time. A virtual machine assignment platform may then execute the node failure prediction algorithm to calculate a node failure probability score for each computing node based on the information in the active node data store. As a result, a virtual machine may be assigned to a selected computing node based at least in part on node failure probability scores.

Computing node failure and health prediction for cloud-based data center

Automatically identifying insights from a dataset and presenting the insights graphically and in natural language text ranked by importance is provided. Different data types and structures in the dataset are automatic recognized and matched with a corresponding specific analysis type. The data is analyzed according to the determined corresponding analysis types, and insights from the analysis are automatically identified. The insights within a given insight type and between insight types are ranked and presented in order of importance in automatically generate charts that visually describe each insight and in natural language text that describes each insight in such a way that it may be understandable to a general audience who may not have a familiarity with statistics.

Automatic insights for spreadsheets

In some examples, a time-series data set can be analyzed and grouped in a fast and efficient manner. For instance, fast grouping of multiple time-series into clusters can be implemented through data reduction, determining cluster population, and fast matching by locality sensitive hashing. In some situations, a user can select a level of granularity for grouping time-series into clusters, which can involve trade-offs between the number of clusters and the maximum distance between two time-series in a cluster.

Fast grouping of time series

Techniques for obtaining a reaction of an individual to content and outputting the reaction of the individual with the content are described herein. The techniques may include obtaining the reaction of the individual as the content is displayed to the individual. The reaction may be obtained by capturing a video of the individual. The reaction of the individual and the content may be output at a same time to another individual. These techniques may also include obtaining the reaction of another individual as the content and the reaction of the individual are output to another individual.

Consumption of content with reactions of an individual

The disclosure generally relates to code recommendation. In one embodiment, code snippets may be extracted using an invocation-centered code slicing process and then grouped according to the method usages. For each method usage, a representative code snippet may be selected and stored in the knowledge base in association with metadata. In operation, the programming context may be obtained and used to query the knowledge base to retrieve one or more code snippets for recommendation. In one embodiment, the recommended code snippets may be ranked to improve the utility and user friendliness, and the metadata may be used to provide variation points and possibly other auxiliary information to improve the operation efficiency and user experience.

Code recommendation

Growing community of inventors

Beijing, China

Yingnong Dang