Multi-level parallelism of process execution in a mutual exclusion domain of a processing system
    23.
    发明授权
    Multi-level parallelism of process execution in a mutual exclusion domain of a processing system 有权
    处理系统的互斥域中的进程执行的多级并行性

    公开(公告)号:US08627331B1

    公开(公告)日:2014-01-07

    申请号:US12771988

    申请日:2010-04-30

    IPC分类号: G06F9/46

    摘要: A technique is described for improving throughput in a processing system, such as a network storage server. The technique provides multiple levels (e.g., a hierarchy) of parallelism of process execution within a single mutual exclusion domain, in a manner which allows certain operations on metadata to be parallelized as well as certain operations on user data. The specific parallelization scheme used in any given embodiment is based at least partly on the underlying metadata structures used by the processing system. Consequently, a high degree of parallelization possible, which improves the throughput of the processing system.

    摘要翻译: 描述了一种用于提高诸如网络存储服务器的处理系统中的吞吐量的技术。 该技术以允许对元数据进行某些操作并行化以及对用户数据的某些操作的方式在单个互斥域内提供过程执行的并行性的多个级别(例如,层级)。 在任何给定实施例中使用的特定并行化方案至少部分地基于处理系统使用的底层元数据结构。 因此,可以进行高度并行化,这提高了处理系统的吞吐量。

    Method and system for replicating stored information
    24.
    发明授权
    Method and system for replicating stored information 有权
    用于复制存储信息的方法和系统

    公开(公告)号:US08595457B1

    公开(公告)日:2013-11-26

    申请号:US12261315

    申请日:2008-10-30

    CPC分类号: G06F13/28 G06F11/1662

    摘要: Method and system for replicating a storage volume is provided. Information is adaptively replicated in a swap mode or a copy mode. When information is copied from a storage volume to a memory buffer, an application determines if another information transfer from the same source volume is pending. If a transfer from the same source is pending, then information is copied from the memory buffer to a stolen buffer in a copy mode. If a transfer from the same source is not pending, then instead of copying the information, the application enables a swap mode. During the swap mode, an operating system for a storage system swaps a pointer from the stolen buffer to information stored in the memory buffer. The memory buffer itself is invalidated so that no other module can access the memory buffer. Because the pointers are swapped, the application accesses information directly from the memory buffer.

    摘要翻译: 提供了复制存储卷的方法和系统。 信息在交换模式或复制模式下自适应地复制。 当信息从存储卷复制到存储缓冲区时,应用程序确定来自同一源卷的另一个信息传输是否处于待处理状态。 如果来自同一个源的传输是挂起的,那么在复制模式下将信息从内存缓冲区复制到被盗的缓冲区。 如果来自同一来源的传输未挂起,则不需要复制信息,应用程序将启用交换模式。 在交换模式期间,用于存储系统的操作系统将指示器从被盗的缓冲器交换到存储在存储器缓冲器中的信息。 内存缓冲区本身无效,因此没有其他模块可以访问内存缓冲区。 因为指针被交换,应用程序直接从内存缓冲区访问信息。

    Efficient file system scan for shared data blocks
    25.
    发明授权
    Efficient file system scan for shared data blocks 有权
    高效的文件系统扫描共享数据块

    公开(公告)号:US08589447B1

    公开(公告)日:2013-11-19

    申请号:US12498980

    申请日:2009-07-07

    IPC分类号: G06F15/16 G06F17/20

    摘要: A technique for efficiently scanning various datasets in a storage system to generate metadata for shared data blocks is provided. In one embodiment, data containers in an active file system of a storage object are first scanned. Remaining images (e.g., snapshots of the file system) are arranged in a chronological order based on the age of such images in the storage object. In some instances, to perform a metadata scan of a given image, the storage server compares each data container in the given image to a corresponding data container in a prior image (from the chronological order) to identify any non-identical data blocks (i.e., blocks that are not common between the data containers of the two images) in each data container. The metadata scanner then generates metadata only for the non-identical data blocks in each data container and skips metadata generation for any remaining data blocks.

    摘要翻译: 提供了一种用于有效扫描存储系统中的各种数据集以产生共享数据块的元数据的技术。 在一个实施例中,首先扫描存储对象的活动文件系统中的数据容器。 剩余图像(例如,文件系统的快照)基于存储对象中的这些图像的年龄按时间顺序排列。 在某些情况下,为了执行给定图像的元数据扫描,存储服务器将给定图像中的每个数据容器与现有图像中的相应数据容器(从时间顺序)进行比较,以识别任何不相同的数据块(即 ,两个图像的数据容器之间不常见的块)。 然后,元数据扫描器仅为每个数据容器中的不相同的数据块生成元数据,并跳过剩余数据块的元数据生成。

    Retention of active data stored in memory
    26.
    发明授权
    Retention of active data stored in memory 有权
    保留存储在内存中的活动数据

    公开(公告)号:US07562078B1

    公开(公告)日:2009-07-14

    申请号:US11394431

    申请日:2006-04-01

    IPC分类号: G06F7/00

    摘要: A method and apparatus for retention of active data stored in memory using multiple indexing systems for data storage. An embodiment of a method for retention of active data in a storage server includes reading data into a first location of a main memory of the storage server. The data in the first location indexes data elements in a long-term data storage in a first manner. The method further provides for copying the data from the first location into a second location in the main memory of the storage server, where the data in the second location indexing the data elements in the long-term data storage in a second manner.

    摘要翻译: 一种用于使用多个索引系统保存存储在存储器中的活动数据用于数据存储的方法和装置。 用于将活动数据保留在存储服务器中的方法的实施例包括将数据读入存储服务器的主存储器的第一位置。 第一位置中的数据以第一方式索引长期数据存储器中的数据元素。 该方法还提供将数据从第一位置复制到存储服务器的主存储器中的第二位置,其中第二位置中的数据以第二方式索引长期数据存储器中的数据元素。