Fast dependency mining using access patterns in a storage system
    11.
    发明授权
    Fast dependency mining using access patterns in a storage system 有权
    使用存储系统中访问模式的快速依赖关系挖掘

    公开(公告)号:US09275117B1

    公开(公告)日:2016-03-01

    申请号:US13706487

    申请日:2012-12-06

    IPC分类号: G06F17/30

    摘要: A system and method is provided in which access patterns of data blocks are used to predict future accesses to the data blocks. One of the types of patterns that may be used in connection with the system described herein is depended blocks. Dependent blocks are blocks that are often referenced one after another in a short period of time. For example, one block (block b) is commonly referenced after another block (block a) a noted time period. The system described herein provides a block dependency algorithm that enables a search task involving searching of block dependencies with respect to proposed dependency definitions. In an embodiment, the search task may be performed based on using of a branch and bound methodology. A search tree is built in which a set of activity vectors (V) are ordered, and a search algorithm is used to traverse and “prune” branches of the tree.

    摘要翻译: 提供了一种系统和方法,其中使用数据块的访问模式来预测将来对数据块的访问。 可以与本文所述系统结合使用的模式类型之一是依赖块。 依赖块是在短时间内经常引用的块。 例如,一个块(块b)通常在另一个块(块a)之后引用指定的时间段。 本文描述的系统提供块依赖性算法,其使得能够涉及关于提出的依赖关系定义搜索块依赖性的搜索任务。 在一个实施例中,可以基于使用分支和绑定方法来执行搜索任务。 构建一个搜索树,其中排列一组活动向量(V),并且使用搜索算法遍历并“修剪”树的分支。