Microsoft Technology Licensing, LLC

Hongqiang Liu

Jitendra Dattatraya Padhye

Yibo Zhu

Ming Zhang

Srikanth Kandula

Ratul Mahajan

Anirudh Badam

Shachar Raindel

Daehyeok Kim

Xin Jin

Jianshu Chen

Distributed computing systems, devices, and associated methods of virtual RDMA switching are disclosed herein. In one embodiment, a method includes intercepting a command from an application in a container to establish an RDMA connection with a remote container on a virtual network. In response to the intercepted command, an RDMA endpoint at a physical NIC of a server is created. The method can also include intercepting another command to pair with a remote RDMA endpoint corresponding to the remote container. The intercepted another command contains data representing a routable network address of the remote RDMA endpoint in the RDMA computer network. Then, the RDMA endpoint created at the physical NIC of the server can be paired with the remote RDMA endpoint using the routable network address of the remote RDMA endpoint.

Virtual RDMA switching for containerized

Virtual RDMA switching for containerized applications

Distributed storage systems, devices, and associated methods of data replication are disclosed herein. In one embodiment, a server in a distributed storage system is configured to write, with an RDMA enabled NIC, a block of data from a memory of the server to a memory at another server via an RDMA network. Upon completion of writing the block of data to the another server, the server can also send metadata representing a memory location and a data size of the written block of data in the memory of the another server via the RDMA network. The sent metadata is to be written into a memory location containing data representing a memory descriptor that is a part of a data structure representing a pre-posted work request configured to write a copy of the block of data from the another server to an additional server via the RDMA network.

Group-based data replication in multi-tenant storage systems

The techniques and/or systems described herein are configured to determine a set of update operations to transition a network from an observed network state to a target network state and to generate an update dependency graph used to dynamically schedule the set of update operations based on constraint(s) defined to ensure reliability of the network during the transition. The techniques and/or systems dynamically schedule the set of update operations based on feedback. For example, the feedback may include an indication that a previously scheduled update operation has been delayed, has failed, or has been successfully completed.

Dynamic scheduling of network updates

Provided are methods and systems for facilitating selection of a cloud configuration for deploying an application program with high accuracy, low overhead, and automatic adaptivity to a broad spectrum of applications and cloud configurations. The methods and systems are designed for building a performance model of cloud configurations, where the performance model is capable of distinguishing an optimal cloud configuration or a near-optimal cloud configuration from other possible configurations, but without requiring the model to be accurate for every cloud configuration. By tolerating the inaccuracy of the model for some configurations (but keeping the accuracy of the final result) it is possible to achieve both low overhead and automatic adaptivity: only a small number of samples may be needed and there is no need to embed application-specific insights into the modeling.

Systems and methods for identifying cloud configurations

Systems, methods, and computer-executable instructions for emulating a network are disclosed. A network topology and an indication of devices to emulate are received. A safe static boundary within the network that comprises each device in the indication of devices to emulate is determined. The safe static boundary is based upon the network topology. For each device to emulate, a unified container and a device container are instantiated. An emulated device is based on the unified container and the device container. Data links between the emulated devices are created based upon the network topology. An emulated network is created using the emulated devices.

High fidelity network emulation

The techniques and/or systems described herein implement a fault handling service that is able to ensure that at least part of a network can avoid congestion (e.g., a link exceeding capacity) as long as a predetermined maximum number of faults is not exceeded. The fault handling service models different combinations of possible faults based on network topology and then computes an amount of traffic to be communicated via individual paths such that congestion is avoided as long as a number of actual faults that occur is less than or equal to the predetermined maximum number of faults.

Growing community of inventors

Woodinville, WA, United States of America

Hongqiang Liu