Microsoft Technology Licensing, LLC

Osman Nuri Ertugay

David G Thaler

Srinivas Raghu Gatta

Dmitry A Anipko

Poornananda R Gaddehosur

Kamalavasan Srinivasan

Andrew J Ritz

Jeroen Vanturennout

Mohammad Saifur Rahman

Keith Edgar Horton

Christian F Huitema

Alireza Dabagh

Gerardo Diaz-Cuellar

Praveen Balasubramanian

Omar Cardona

Shankar Seal

Joseph Nievelt

Nicholas David Wood

Dinesh Kumar Govindasamy

Poornananda Gaddehosur Ramachandra

Daniel Havey

Mahender Hari

Anirban Paul

Embodiments relate to hypervisors that provide hardware isolated virtualization environments (HIVEs) such as containers and virtual machines (VMs). A first HIVE includes a first virtual network interface card (NIC) and a second HIVE includes a second virtual NIC. Both virtual NICs are backed by the same physical NIC. The physical NIC has an Internet Protocol (IP) address. The virtual NICs are assigned the same IP address as the physical NIC. A networking stack of the hypervisor receives inbound packets addressed to the IP address. The networking stack steers the inbound packets to the virtual NICs according to tuples of the inbound packets. Packets emitted by the virtual NICs comprise the IP address, pass through the network stack, and are transmitted by the physical NIC with headers comprising the IP address.

Hypervisor flow steering for address sharing

The technologies disclosed herein provide improvements to the Low Extra Delay Background Transport (LEDBAT) protocol. Some aspects of the present disclosure introduce an adaptive congestion window gain for background connections. In some configurations, a gain value for influencing the transmission rate of a background connection is dynamically adjusted based on data indicating a round trip time (RTT). The RTT includes a sum of a time in which the data is communicated to a remote device and a time in which acknowledgement is data returned from the remote device. In some configurations, the gain is decreased when the RTT is below a threshold and the gain is increased when the RTT is above the threshold. Among other features, the present disclosure also provides techniques involving a modified slow-start, multiplicative decrease and periodic slowdowns. The features disclosed herein mitigate some existing issues, such as latency drift, inter-LEDBAT fairness, and unnecessary slowdowns.

Adaptive gain reduction for background connections

Keep alive management techniques are described. In one or more implementations, a keep alive interval is calculated by an operating system of the computing device. The keep alive interval is used to maintain one or more notification channels between one or more applications of the computing device and a network.

Keep alive management

Wake pattern management techniques are described. In one or more implementations, network traffic received by a network interface device of a computing device is monitored and a traffic pattern is recognized in the monitored network traffic. An application of the computing device is identified that corresponds to the recognized traffic pattern and responsive to this identification, at least a portion of the identified application is woken.

Wake pattern management

A network interface that connects a computing device to a network may be configured to process incoming packets and determine an action to take with respect to each packet, thus decreasing processing demands on a processor of the computing device. The action may be indicating the packet to an operating system of the computing device immediately, storing the packet in a queue of one or more queues or discarding the packet. When the processor is interrupted, multiple packets aggregated on the network interface may be indicated to the operating system all at once to increase the device's power efficiency. Hardware of the network interface may be programmed to process the packets using filter criteria specified by the operating system based on information gathered by the operating system, such as firewall rules.

Mechanism to save system power using packet filtering by network interface

In embodiments of a programming interface for data communications, a request queue and a completion queue can be allocated from a user-mode virtual memory buffer that corresponds to an application. The request queue and the completion queue can be pinned to physical memory and then mapped to kernel-mode system addresses so that the request queue and the completion queue can be accessed by a kernel-mode execution thread. A request can be received from an application for the kernel to handle data in the request queue, and a system issued to the kernel for the kernel-mode execution thread to handle the request. The kernel-mode execution thread can then handle additional requests from the application without additional system calls being issued.

Programming interface for data communications

Growing community of inventors

Bellevue, WA, United States of America

Osman Nuri Ertugay