Packet dispersion method based on network adapter

2013 
The invention provides a packet dispersion method based on a network adapter. The packet dispersion method based on the network adapter is applied to a network server; and the method comprises the following steps: dividing a share memory into a plurality of memory blocks by a server control core based on a memory division algorithm; and sending a share memory base address, memory block numbers and memory block sizes to the network adapter; after receiving a packet from outside by the network adapter, determining the number of a network adapter data core transmitting the packet and the number of a corresponding memory block storing the packet by a network adapter control core based on a scheduling algorithm, and sending the memory block number to the corresponding network adapter data core; after receiving the memory block number from the network adapter control core by the network adapter data core, calculating a memory block address storing the packet based on the share memory base address, the memory block numbers and the memory block sizes, and storing the packet into the memory block. By the packet dispersion method based on the network adapter, the utilization ratio of a network adapter multi-core CPU in the network server can be effectively increased, accordingly avoiding the resource waste of the network adapter multi-core CPU, and achieving a real parallel access to the share memory by the network adapter multi-core CPU.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []