International Business Machines Corporation

Xiao Yu Wang

Da Lei Zhang

Chen Zhang

Yu Chen Zhou

Li Yang

Wen Bao Yin

Xiang Liu

Long Wen Lan

Li Li Gu

Wen Wu Na

Jia Jun Lu

Le Yi Zhou

Lan Ying Shen

Jie Zou

Aspects of the present invention disclose a method for a two-node storage system. The method includes one or more processors creating a plurality of first logic unit groups in a first storage node of a storage system. The method further includes mapping each of the plurality of first logic unit groups to a number of storage slices from different storage devices in the first storage node. The method further creating a plurality of second logic unit groups in a second storage node of the storage system, by mirroring storage slices from a storage device in the first storage node to multiple storage devices in the second storage node. In response to identifying a failure of a first storage device in the first storage node, the method further includes recovering lost data based on data in the second storage node.

Two-node high availability storage system

A method and a system for creating a workload distribution plan on a storage system. The method includes, for each storage location, initializing components for at least one data type, such that the components establish a baseline distribution percentage. The method also includes writing data packages onto the storage system over a predetermined time period. The method also includes, for each data type on each storage location, creating a workload distribution plan, where the workload distribution plan distributes a workload of the storage system to the storage location. The method also includes implementing the workload distribution plan on the storage system.

Dynamically altered data distribution workload on a storage system

A batch of input/output (I/O) requests, directed to multiple physical domains of storage area network disk (SAN disk), are prioritized for processing based on the workload (pressure) of the target physical domains, and further based on the workload of multiple input/output (I/O) paths available for transmission of the I/O data to or from the physical domains. Requests directed to a physical domain with low pressure (relative to other physical domains to which some of the I/O requests are directed) are queued up and sent to the physical domain on an I/O path having a low workload (relative to other available I/O paths). Requests directed to a physical domain under relatively high pressure are queued up and sent to the physical domain on an I/O path having a relatively higher workload.

Parallel input/output via multipath software

Growing community of inventors

Shanghai, China

Xiao Yu Wang