Microsoft Technology Licensing, LLC

Linkedin Corporation

Workday, Inc.

Anant Rao

Sajid Topiwala

Pritesh Shah

Arunprasad Venkatraman

Walter Scott Johnson

Haricharan K Ramachandra

Kishore Kumar Pusukuri

Prasanna Vijayanathan

Sreedhar Veeravalli

Arman H Boehm

Mark E Pascual

Jui Ting Weng

A system for outlier detection and removal comprises an interface and a processor. The interface is configured to receive a data set. The processor is configured to determine a cleaned data set by removing outliers, wherein determining the cleaned data set comprises determining a type of distribution, in response to the type of distribution being normal, determining the outliers using covariance estimation, in response to the type of distribution not being normal, determining the outliers using density based clustering, and determining the cleaned data set by removing the outliers from the data set, determine a coefficient of variation of the cleaned data set, determine whether the coefficient of variation is greater than a threshold coefficient of variation, and in response to the coefficient of variation being greater than the threshold coefficient of variation, determine a new cleaned data set by removing a new set of outliers from the cleaned data set.

Outlier identification and removal

A system and method for generating a simulation environment are provided. In example aspects, in response to a request from a user of an environment simulation system, specifying a desired web performance metric to result from the execution of a particular content item, the environment simulation system constructs a simulation environment. The environment simulation system is configured to test the particular content item within the constructed simulation environment and compare a baseline web performance metric to the requested web performance metric. In some aspects, if the requested web performance metric is not achieved, the environmental simulation system is configured to perform tuning and continue simulating until the requested web performance metric is achieved. Once tuning is complete, a user can modify a content item and continue simulating to observe improvements in web performance of a content item for a user in a particular geographic region.

Environment simulator for user percentile

The disclosed embodiments provide a system for processing data. During operation, the system obtains an attribute of a stack trace of a software program. Next, the system uses the attribute to select an address-translation instance from a set of address-translation instances for processing the stack trace. The system then provides the stack trace to the selected address-translation instance for use in translating a set of memory addresses in the stack trace into a set of symbols of instructions stored at the memory addresses.

Efficient address-to-symbol translation of stack traces in software programs

Techniques are provided for recording service invocation traffic in a format that is suitable for subsequent replay, perhaps in a different environment. In one technique, a computing device records invocation traffic. The computing device receives a first request that is directed to a first service. In response to receiving the first request, one or more computers store an identifier of the first request into a second request. After storing the identifier into the second request, the computing device causes the second request to be sent to a second service. After causing the second request to be sent to the second service, the computing device receives, from the second service, a response that contains the identifier of the first request. The one or more computers store, in a single record, the first request, the second request, and the response.

Replay-suitable trace recording by service container

Techniques are provided for mocking downstream services while regression testing a service. In one technique a version of a service processes a request, during which one or more computers intercept downstream service calls made by the version of the service. For each downstream service call of the downstream calls, a computer provides, to the version of the service, a response of the downstream service call, which was previously recorded while processing the request in a production environment. Processing, by the version of the service, the request involves processing the request based on the response of at least one of the downstream service calls.

Regression testing of software services

Growing community of inventors

Fremont, CA, United States of America

Anant Rao