Load-balancing structure for packet switches and its constructing method
    1.
    发明授权
    Load-balancing structure for packet switches and its constructing method 有权
    分组交换机的负载均衡结构及其构建方法

    公开(公告)号:US08902887B2

    公开(公告)日:2014-12-02

    申请号:US12739729

    申请日:2009-10-31

    CPC分类号: H04L49/101 H04L49/552

    摘要: This invention provides a load-balancing structure for packet switches and its constructing method. In this method, the structure based on self-routing concentrators is divided into two stages, that is, a first stage and a second stage fabric. A virtual output group queue (VOGQ) is appended to each input group port of the first stage fabric, and a reordering buffer (RB) is configured behind each output group port of the second stage fabric. Packets stored in the VOGQ are combined into data blocks with preset length, which is divided into data slices of fixed size, finally each data slice is added an address tag and is delivered to the first stage fabric for self-routing. Once reaching the RB, data slices are recombined into data blocks. This invention solves the packet out-of-sequence problem in the load-balancing Birkhoff-von Neumann switching structure and improves the end-to-end throughput.

    摘要翻译: 本发明提供了分组交换机的负载平衡结构及其构造方法。 在这种方法中,基于自路由集中器的结构分为两个阶段,即第一阶段和第二阶段结构。 将虚拟输出组队列(VOGQ)附加到第一级结构的每个输入组端口,并且在第二级结构的每个输出组端口之后配置重新排序缓冲器(RB)。 存储在VOGQ中的数据包被组合成具有预设长度的数据块,其被划分为固定大小的数据片段,最后每个数据片段添加地址标签,并且被递送到第一级结构以进行自路由。 一旦到达RB,数据切片被重新组合成数据块。 本发明解决了负载平衡Birkhoff-von Neumann交换结构中的分组失序问题,提高了端到端吞吐量。

    Load-Balancing Structure for Packet Switches and Its Constructing Method
    2.
    发明申请
    Load-Balancing Structure for Packet Switches and Its Constructing Method 有权
    分组交换机的负载平衡结构及其构建方法

    公开(公告)号:US20110176425A1

    公开(公告)日:2011-07-21

    申请号:US12739729

    申请日:2009-10-31

    IPC分类号: H04L12/26

    CPC分类号: H04L49/101 H04L49/552

    摘要: This invention provides a load-balancing structure for packet switches and its constructing method. In this method, the structure based on self-routing concentrators is divided into two stages, that is, a first stage and a second stage fabric. A virtual output group queue (VOGQ) is appended to each input group port of the first stage fabric, and a reordering buffer (RB) is configured behind each output group port of the second stage fabric. Packets stored in the VOGQ are combined into data blocks with preset length, which is divided into data slices of fixed size, finally each data slice is added an address tag and is delivered to the first stage fabric for self-routing. Once reaching the RB, data slices are recombined into data blocks. This invention solves the packet out-of-sequence problem in the load-balancing Birkhoff-von Neumann switching structure and improves the end-to-end throughput.

    摘要翻译: 本发明提供了分组交换机的负载平衡结构及其构造方法。 在这种方法中,基于自路由集中器的结构分为两个阶段,即第一阶段和第二阶段结构。 将虚拟输出组队列(VOGQ)附加到第一级结构的每个输入组端口,并且在第二级结构的每个输出组端口之后配置重新排序缓冲器(RB)。 存储在VOGQ中的数据包被组合成具有预设长度的数据块,其被划分为固定大小的数据片段,最后每个数据片段添加地址标签,并且被递送到第一级结构以进行自路由。 一旦到达RB,数据切片被重新组合成数据块。 本发明解决了负载平衡Birkhoff-von Neumann交换结构中的分组失序问题,提高了端到端吞吐量。