Microsoft Technology Licensing, LLC

Matthew Jeremiah Eason

Baskar Sridharan

Andrew J Lagattuta

John R Douceur

Edmund Bernard Nightingale

Landy Wang

Michael G Myrah

Garrett Leischner

Asad Yaqoob

Stephan Doll

Paul Luo Li

Abdelsalam Heddaya

Chandan Aggarwal

Akil M Merchant

Josh David McKone

A system is described that can detect pressure anomalies within a data center, generate an alert when anomalies are detected, and initiate remediative actions. The system monitors each of a plurality of fans used to dissipate heat generated by one or more servers to obtain data that indicates how an actual speed of each of the fans relates to a target speed thereof. The system compares the obtained data to reference data that indicates, for each of the plurality of fans, how an actual speed of the fan relates to a target speed thereof in a substantially pressure-neutral environment. Based on the comparison, the system determines whether or not a pressure anomaly exists. If the system determines that a pressure anomaly exists, then the system may perform various actions such as generating an alert and modifying a manner of operation of one or more of the fans or servers.

Data center pressure anomaly detection and remediation

To prevent overloading of a shared secondary power feed, active transfer switches are controlled to switch discrete hardware from the shared secondary to a corresponding primary power feed, even if such a primary is currently nonoperational thereby resulting in deactivation of such hardware. A threshold amount of power draw from the shared secondary is established, such as to ensure that failure of a single primary can be accommodated, and active transfer switches are controlled so as to prevent such a threshold from being exceeded. Alternatively, or in addition, active transfer switches are controlled such that they, in the event of an interruption in the power being provided by the primary power feed, either switch to the shared secondary or, in accordance with the aforementioned control, do not switch despite the interruption in power.

Active power transfer switch control to reduce service impacts due to electrical power interruptions

Power monitoring in a server set that has a shared power supply system that includes multiple power supply unit sets that each provide power to the server set. A controller monitors power usage of the server set, and evaluates the monitored power usage against the power supply capability of each of the power supply unit sets. Based on the evaluation, the controller can detect when a particular power supply unit set is no longer capable of solely providing all of the power that the server set is consuming or anticipated to consume, and then reduce power accordingly so that the particular power supply unit set could provide such power.

Monitoring of shared server set power supply units

Automated health monitoring and recovery is provided for infrastructure devices supporting server devices in a data center. Health analysis operations may be selected to be performed on an infrastructure device based on the capabilities of the infrastructure device and/or how the infrastructure device is being used to support server devices in the data center. If the infrastructure device is unhealthy, an automated recovery operation may be performed. The automated recovery operation may include recovery actions selected based on the capabilities of the infrastructure device, the failure mode of the infrastructure device, and/or how the infrastructure device is being used to support server devices in the data center.

Health monitoring and recovery for infrastructure devices

A system and method for mitigating memory errors in a computer system. Faulty memory is identified and tested by a memory manager of an operating system. The memory manager may perform diagnostic tests while the operating system is executing on the computer system. Regions of memory that are being used by software components of the computer system may also be tested. The memory manager maintains a stored information about faulty memory regions. Regions are added to the stored information when they are determined to be faulty by a diagnostic test tool. Memory regions are allocated to software components by the memory manager after checking the stored information about faulty memory regions. This ensures a faulty memory region is never allocated to a software component of the computer system.

Systematic mitigation of memory errors

A method of detecting a network hang and restoring an application that communicates on a connection giving rise to the network hang. A user experience may be improved by providing the user with an option to restore the hung application without losing unsaved data or state information. The network hang may be detected when the user tries to terminate the application. The method may include determining whether the network hang is recoverable, which may involve diagnosing a type of the hang. If recoverable, a network connection reset mechanism may be triggered by instructing a network stack of the computer to terminate the network connection.

Growing community of inventors

Sammamish, WA, United States of America

Matthew Jeremiah Eason