Future Wei Technologies, Inc.

Yunsong Lu

Wenzhe Zhou

Yan Chen

Yan Sun

A method includes an interconnect performing a partial direct memory access of a packet to obtain packet header data of a packet header, using the packet header data to generate a packet forwarding decision identifying a packet destination address, and initiating a direct memory access of a packet payload to the packet destination address.

Partially deferred packet access

A virtual switch for packet switching includes an ingress traffic steering manager executing on circuitry and coupled to receive packets from multiple virtual machines or containers, multiple data plane providers, each data plane provider having a data path coupled to selectively receive the packets from the ingress traffic steering manager, and wherein the ingress traffic steering manager classifies the received packets and selects available data paths based on the classification of the packets and a set of distribution rules.

Multi-path virtual switching

The disclosure relates to technology for supporting a virtual switch to change data plane providers on a framework supporting multiple data plane providers. A processing device receives a request to change a first data plane provider, where the virtual switch is configured with a topology on the first data plane provider to use a flow management protocol. The virtual switch includes network interfaces connected to ports to enable communication among entities attached to the network interfaces by forwarding data packets within a first datapath of the first data plane. In response to the change, the network interfaces are disconnected, the first datapath is removed and a second datapath is created. The virtual switch is then configured to operate with the second datapath while retaining the flow management protocol and the topology, such that the entities communicate by forwarding data packets within the second datapath on the second data plane.

Multiple provider framework for virtual switch data planes and data plane migration

The disclosure relates to technology performing a cyclic redundancy check (CRC). Data is divided into a plurality of blocks, each of the plurality of blocks having a fixed size equal to a degree of a generator polynomial. A CRC computation is independently performed on each of the plurality of blocks, and the CRC computation for each of the plurality of blocks is combined by application of an exclusive or (XOR) operation.

Cyclic redundancy check calculation for multiple blocks of a message

An embodiment method of network protocol offloading for a packet traveling along a path having a plurality of hops in a virtual network includes checking, for each of the plurality of hops, that a virtual receiver node is safe for network protocol offloading and, when safe, deferring a network protocol processing task, forwarding the packet to a next hop among the plurality of hops, executing the network protocol processing task at a first earliest hop along the path where the virtual receiver node is unsafe for network protocol offloading, and executing the network protocol processing task at a second earliest hop along the path where the virtual receiver node includes a physical network interface that is safe for network protocol offloading.

System and method for network protocol offloading in virtual networks

A method and apparatus are disclosed for network resource virtual partitioning. An embodiment method includes mapping a plurality of hardware functions at a plurality of physical network interface devices into a plurality of virtual partitions (VPs) implemented using software, wherein the VPs are configured to manage and operate independent from one another the corresponding hardware functions at the physical network interface devises. An embodiment apparatus includes a processor configured to aggregate a plurality of hardware functions at a plurality of physical network interface devices into a plurality of virtual partition aggregations (VPAs), wherein the VPAs are configured to manage and operate independent from one another a plurality of corresponding subsets of the hardware functions to serve one or more clients.

Method and apparatus for network resource virtual partitioning

Disclosed herein are a system and method for network protocol offloading in virtual networks. An embodiment method of offloading a network protocol processing task for a message includes receiving, at a virtual switch, the message from a virtual source node destined for a virtual receiver node. The network protocol processing for the message is offloaded from the virtual source node and deferred when the virtual receiver node is safe for network protocol offloading. When the virtual receiver node is unsafe for network protocol offloading, the network protocol processing task is executed within the virtual switch. When the virtual receiver node is safe for network protocol offloading and a path between the virtual switch and the virtual receiver node includes a physical network interface, the network protocol processing task is executed using the physical network interface.

An apparatus for sharing network capabilities of a host system by one or more virtual machines. A unified virtual network interface instance is included within an operating system of a virtual machine operating within a virtual environment. The interface instance provides device specific network capability for the virtual machine. A host system supports the virtual machine and includes a network capability provider instance that interfaces with the unified virtual network interface to provide the device specific network capability for the virtual machine. A network capability provider enforces sharing of host-wide network capability as implemented through the one or more unified virtual network interfaces supported by the host system.

Growing community of inventors

Mountain View, CA, United States of America

Yunsong Lu