Microsoft Technology Licensing, LLC

Webtv Networks, Inc.

Eric Lynn Eilebrecht

Stephen Harris Toub

Vance Palmer Morrison

Maneesh Sah

Yoram Bernet

Shreedhar Madhavapeddi

Marcus Jon Jager

Ofer Bar

Karthick Krishnamoorthy

Steven Robert Sandke

Matthew Terence Hayes

Anastasios Kasiolas

Marc Andrew Power

Gregory Paperin

John Joseph Duffy

Joseph L Hellerstein

Paul F Ringseth

Mads Torgersen

Rajesh Sundaram

Kumar Gaurav Khanna

Ladislav Prosek

Charles Reeves Little

Diana Milirud

Melitta L G Andersen

Rajesh Sundram

Erika Fuentes

An example method of parsing data in a distributed computing system includes assigning a worker k to scan a data chunk k at a data storage node at which data chunk k is stored, each of one or more data chunks including a partial record, receiving, from each worker k, a number of instances of a context-varying symbol, a position in the data chunk k of a first instance of a context-dependent symbol after an even number of instances of the context-varying symbol, and a position in the data chunk k of a first instance of the context-dependent symbol after an odd number of instances of the context-varying symbol, for each data chunk that includes a partial record, computing an adjusted data chunk including data from a later data chunk to complete the partial record, parsing the adjusted data chunks and executing a query against the plurality of records.

Parallelized parsing of data in cloud storage

Embodiments of the present disclosure are directed a distributed storage system that supports query data. One embodiment provides, at a front-end node, a method comprising receiving a query regarding data stored in an object, obtaining information identifying data nodes at which blocks of the object are stored, sending, to a first data node, a query regarding data in a first block, and receiving results for the query regarding the data in the first block comprising an incomplete record. The method further comprises sending, to a second data node, a query regarding data in a second block comprising the incomplete record, the second block being stored on a different physical device than the first block, and receiving results for the query regarding the data in the second block. The method further comprises outputting the results for the queries regarding the data in the first block and the data in the second block.

Querying data in a distributed storage system

One example provides, at a front-end node of a distributed storage system, a method comprising receiving a query regarding data stored in an object, obtaining information identifying data nodes at which blocks of the object are stored, sending, to a first data node, a query regarding data in a first block, and receiving results for the query regarding the data in the first block comprising an incomplete record. The method further comprises sending, to a second data node, a query regarding data in a second block comprising the incomplete record, the second block being stored on a different physical device than the first block, and receiving results for the query regarding the data in the second block. The method further comprises outputting the results for the queries regarding the data in the first block and the data in the second block.

Computer systems attempt to manage resource pools of a dynamic number of similar resources and work tasks in order to optimize system performance. Work requests are received into the resource pool having a dynamic number of resources instances. An instance-throughput curve is determined that relates a number of resource instances in the resource pool to throughput of the work requests. A slope of a point on the instance-throughput curve is estimated with stochastic gradient approximation. The number of resource instances for the resource pool is selected when the estimated slope of the instance-throughput curve is zero.

Managing pools of dynamic resources

Server-side control of client-side quality-of-service tracking is provided. When a service is being provided to a client device, a server provides aggregation rules to a client device as part of the service. The client device executes instructions in the aggregation rules to collect QoS metrics regarding the service on the client device and to manipulate the QoS metrics to generate aggregated QoS data. The client device reports the aggregated QoS data to a server for storage and analysis.

Rule-based system for client-side quality-of-service tracking and reporting

A software optimization system isolates an effect of a change in a control variable from effects of ongoing, unknown changes in other variables. The system discards effects due to noise so that effects of interest to a programmer are more easily visible. The software optimization system treats variations in one or more control variables and in the output of the system as signals. The system varies the control variable at a specific frequency unlikely to correlate with uncontrolled variations in external variables. The system uses digital signal processing (DSP) techniques to filter the output, isolating the frequency of the control variable variation. The system then compares the resulting filtered output to the input to determine the approximate effect of the variation in the control variable.

Isolating changes in dynamic systems

The present invention extends to methods, systems, and computer program products for memory pinning through buffer encapsulation. Within a managed execution environment, a wrapper object encapsulates a memory buffer that is to be shared with a native routine executing in a native execution environment. The wrapper object manages operation of a memory manager on a memory heap corresponding to the memory buffer. The wrapper object includes a first function which sets a pin on the memory buffer and returns a pointer identifying the memory buffer. Setting the pin causes the memory manager to cease moving the memory buffer within the memory heap. The wrapper object also includes a second function which releases the pin on the memory buffer.

Memory pinning through buffer encapsulation

The present invention extends to methods, systems, and computer program products for handling synchronous operations by means of asynchronous operations. Upon completion of an asynchronous operation, a state flag is accessed. The state flag indicates whether or not a sync-over-async wrapper/adapter requested execution of the asynchronous operation. The sync-over-async wrapper/adapter is currently blocked awaiting notice of completion of the asynchronous operation. Based on the state flag, results of the asynchronous operation are stored at a location accessible by the sync-over-async wrapper. A completion signal is sent to the sync-over-async wrapper.

Async wrapper handling execution of asynchronous operations for synchronous and asynchronous routines

A system and method for handling exceptions in a multi-threaded computing environment. Information, such as that relating to an error state or pertaining to the propagation history of an exception, is stored in a separate object from the exception object itself. The separate propagation information object is accessible to the plurality of threads that are used to execute a user task. The separate object allows rich diagnostic information pertaining to the exception and its propagation through multiple threads to be presented to the developer of the software.

Separation of error information from error propagation information

Server-side control of client-side quality-of-service tracking is provided. When an event occurs at a client, a measurement client collects event parameters and generates aggregated measurement data in accordance with aggregation rules that are provided to the measurement client from a server-side measurement service. The measurement client periodically reports the aggregated measurement data to the measurement service for storage and analysis.

A first in, first out queue uses a sequence of arrays to store elements in the queue. The arrays are constructed using a lock free queue, and within each array, a lock free mechanism may be used to enqueue and dequeue elements. Many embodiments may use atomic operations to ensure successful placement of elements in the queue, as well as remove elements from the queue. The atomic operations may be used within a loop until successful.

Lock free queue

Static data race analysis of at least a portion of a multi-threaded application in order to identify potential data race defects in the multi-threaded application. The static data race analysis includes intra-component static analysis as well as inter-component static analysis. The intra-component static analysis for a given component involves identifying a set of memory accesses operations in the component. For each of at least one of the set of memory access operations, the analysis determines whether there is a data race protection element associated with the memory access command.

Static data race detection and analysis

A method of providing asynchronous programming support at run time for scope-based ambient state at a method call is disclosed. In response to the method call, an asynchronous method stack frame is created along with ambient state for the asynchronous method call. In one example of the method, a current frame of a method stack is captured. A new frame is instantiated by setting a previous reference of the new frame to the current frame and moving data from the previous frame into the current frame. One example of instantiating the new frame includes pushing the current frame onto the stack. In another example, the new frame is lazily created when changes have been made. A state machine used to transform the synchronously-created method into an asynchronous method is moved from a current state to a next state. The captured frame is restored.

Ambient state for asynchronous methods

Aspects of the subject matter described herein relate to semaphores. In aspects, a feedback mechanism is described that uses a semaphore to attempt to maximize throughput of a computer system. Throughput is tracked over periods of time. At or after the end of a period of time, the throughput of the period of time together with other data is used to determine whether a max count for a semaphore is to be changed and by how much. This monitoring of throughput and modifying of the max count may continue over subsequent periods of time to attempt to maximize throughput.

Growing community of inventors

Woodinville, WA, United States of America

Eric Lynn Eilebrecht