公开(公告)号：US20210157676A1

公开(公告)日：2021-05-27

申请号：US16807724

申请日：2020-03-03

Applicant: HITACHI, LTD.

Inventor： Kazushi NAKAGAWA ,  Mitsuo HAYASAKA ,  Yuto KAMO

IPC: G06F11/10 ,  G06F11/30 ,  G06F3/06

Abstract: A distributed storage places data units and parity units constituting a stripe formed by divided data into storage nodes in a distributed manner. In reference to determination formulas, either a full-stripe parity calculation method or an RPM parity calculation method is selected so as to minimize an amount of network traffic.

公开(公告)号：US20190220454A1

公开(公告)日：2019-07-18

申请号：US16304722

申请日：2016-10-20

Applicant: HITACHI, LTD.

Inventor： Yuko MATSUI ,  Mitsuo HAYASAKA ,  Christopher James ASTON ,  Jonathan SMITH ,  James GIBBS ,  Daniel PICKEN ,  Simon CROSLAND ,  Hieu Hanh LE

IPC: G06F16/22 ,  G06F3/06

CPC classification number: G06F16/2272 ,  G06F3/0611 ,  G06F3/064 ,  G06F3/067 ,  G06F3/0689 ,  G06F15/16 ,  G06F15/173 ,  G06F16/2246 ,  H04L12/46 ,  H04L12/4625

Abstract: In a distributed storage of data objects in a cluster system, a data object is distributed across a group of node apparatuses in units of data blocks. For each child object, the metadata structure of the respective child object is stored to one of the node apparatuses of the group of node apparatuses for managing locations of data blocks of the data object and includes a root metadata node and one or more direct metadata nodes, optionally further including one or more indirect metadata nodes; and the metadata structure of the parent object is distributed across the plural node apparatuses of the group of node apparatuses for managing locations of child objects of the data object and includes, on each node apparatus of the group of node apparatuses, a root metadata node and one or more direct metadata nodes, optionally further including one or more indirect metadata nodes.

公开(公告)号：US20190196908A1

公开(公告)日：2019-06-27

申请号：US16327349

申请日：2017-06-15

Applicant: Hitachi, Ltd.

Inventor： Takeru CHIBA ,  Mitsuo HAYASAKA

IPC: G06F11/10 ,  G06F3/06

Abstract: Each of multiple nodes has a processor, a memory, and a storage region. In a process of updating a stripe, the processors generate an intermediate parity from a data block included in a first node and a parity block included in the first node and included in the stripe to be processed, transfer the intermediate parity to a second node, cause the intermediate parity to be stored as a parity in a block of the second node, and configure the stripe with the data block from which the intermediate parity has been generated, the block storing the parity, and a data block included in a node other than the first and second nodes and included in the stripe to be processed.

公开(公告)号：US20190138358A1

公开(公告)日：2019-05-09

申请号：US16180866

申请日：2018-11-05

Applicant: HITACHI, LTD.

Inventor： Kazumasa MATSUBARA ,  Mitsuo HAYASAKA

IPC: G06F9/50

Abstract: A computer system includes a plurality of task processing nodes capable of executing tasks and a task management node which determines which a task processing node to allocate a new task and each task processing nodes includes a memory capable of caching data to be used by an allocation task which is a task allocated to the task processing node. The task management node stores task allocation history information including a correspondence relationship between the allocation task and the respective task processing node A CPU of the task management node determines a degree of similarity between the new task and the allocated task, determines the task processing node to which the new task should be allocated from the task processing nodes included in the task allocation history information based on the degree of similarity, and allocates the new task.

公开(公告)号：US20180373435A1

公开(公告)日：2018-12-27

申请号：US15777782

申请日：2016-02-26

Applicant: Hitachi, Ltd.

Inventor： Mitsuo HAYASAKA ,  Kazumasa MATSUBARA

IPC: G06F3/06

CPC classification number: G06F3/065 ,  G06F3/06 ,  G06F3/061 ,  G06F3/0619 ,  G06F3/064 ,  G06F3/067 ,  G06F13/10 ,  G06F2211/1028

Abstract: A first storage node generates a plurality of distributedly arranged write data blocks from write data and generates a first redundant data block from the plurality of distributedly arranged write data blocks. One distributedly arranged write data block is arranged in a second data block and the first redundant data block is arranged in a third storage node. A second storage node generates a second redundant data block from the plurality of distributedly arranged write data blocks selected from the distributedly arranged write data block held therein. The second storage node rearranges each of the plurality of selected distributedly arranged write data blocks in a rearrangement destination storage node and arranges the second redundant data block in a storage node other than the rearrangement destination storage node.

公开(公告)号：US20150356134A1

公开(公告)日：2015-12-10

申请号：US14760721

申请日：2013-05-13

Applicant: HITACHI, LTD.

Inventor： Mitsuo HAYASAKA ,  Hitoshi KAMEI ,  Atsushi SUTOH

IPC: G06F17/30

CPC classification number: G06F17/30371 ,  G06F3/0608 ,  G06F3/0641 ,  G06F3/067 ,  G06F17/30097 ,  G06F17/30159 ,  G06F17/3033 ,  G06F17/30342

Abstract: Chunk de-duplication performance is improved. A de-duplication system has a cut-out processing unit which inputs a content from a client terminal thereinto, determines a calculation range from a predetermined maximum chunk size and a predetermined minimum chunk size, divides the calculation range into at least two small calculation ranges, sets the positions of windows for rolling hash calculation so that the rolling hash calculation is continuous between the two small calculation ranges, and subjects the at least two small calculation ranges to the rolling hash calculation with shifting of the windows based on parallel processing to cut out a chunk from the content, and a de-duplication processing unit which does not store the cut-out chunk into a storage device when the chunk having the same contents as those of the cut-out chunk is already stored in the storage device.

Abstract translation: 块重复删除性能得到提高。 重复数据删除系统具有从客户终端输入内容的切出处理单元，从预定的最大块大小和预定的最小块大小确定计算范围，将计算范围划分为至少两个小的计算范围 ，设置用于滚动哈希计算的窗口的位置，使得滚动哈希计算在两个小的计算范围之间是连续的，并且基于平行处理切换窗口来移动至少两个小的计算范围到滚动哈希计算 从该内容中取出一个块，以及一个重复数据删除处理单元，当具有与切出的块相同内容的块已经存储在存储设备中时，该存储设备中不存储切出的块。

公开(公告)号：US20240377947A1

公开(公告)日：2024-11-14

申请号：US18583525

申请日：2024-02-21

Applicant: Hitachi, Ltd.

Inventor： Pablo MARTINEZ LERIN ,  Mitsuo HAYASAKA

IPC: G06F3/06

Abstract: A system receives a quota request in which a tenant, one or more locations, and a capacity upper limit related to a requested quota are designated, executes conflict determination n based on quota information, and adds information related to the requested quota to the quota information when a result of the conflict determination is false. The quota information includes information representing a tenant, one or more locations, and a capacity upper limit for each quota of the plurality of storage devices at the plurality of locations. For each of the plurality of locations, a capacity usable by the tenant among a capacity of the storage device at the location is equal to or less than the capacity upper limit of the quota corresponding to the location and the tenant.

公开(公告)号：US20240103934A1

公开(公告)日：2024-03-28

申请号：US18181732

申请日：2023-03-10

Applicant: Hitachi, Ltd.

Inventor： Shimpei NOMURA ,  Mitsuo HAYASAKA

IPC: G06F9/50

CPC classification number: G06F9/5094

Abstract: A memory of an application platform stores per site a performance model indicating a relationship between program performance and a resource amount of hardware necessary for realizing program performance, and an electric power consumption model indicating a relationship between a resource allocation amount that is an amount allocated to the program, and an electric power consumption amount consumed when the program is executed. The CPU receives target performance information that indicates target performance for the program, calculates per site a necessary allocation amount and a necessary electric power consumption amount that are a resource allocation amount and an electric power consumption amount necessary for realizing the target performance by using the target performance information, the performance model, and the electric power consumption model, and creates a container/data allocation plan that is an allocation plan of an execution platform of the program and data based on a result of the calculation.

公开(公告)号：US20230334072A1

公开(公告)日：2023-10-19

申请号：US18170614

申请日：2023-02-17

Applicant: Hitachi, Ltd.

Inventor： Keiichi MATSUZAWA ,  Mitsuo HAYASAKA

IPC: G06F16/332 ,  G06F16/33

CPC classification number: G06F16/3329 ,  G06F16/3326 ,  G06F16/3349

Abstract: A customer inputs a question sentence indicating a problem that the customer needs to resolve, to an automatic question answering system, and the system answers the question sentence. A history of the conversation is recorded in the system as conversation history data. When the system fails to give a suitable answer in a question-and-answer session, the system escalates the question to a support representative. In such a case, the question sentences and an answer sentence given by the support representative to resolve the problem are added to question-answer pair data as new question-answer pairs. The accuracy of automatic question answering is thus enhanced.

公开(公告)号：US20230281161A1

公开(公告)日：2023-09-07

申请号：US17901340

申请日：2022-09-01

Applicant: Hitachi, Ltd.

Inventor： Takeshi KITAMURA ,  Ryo FURUHASHI ,  Mitsuo HAYASAKA ,  Shimpei NOMURA ,  Masanori TAKATA

IPC: G06F16/11 ,  G06F16/182 ,  G06F16/188

CPC classification number: G06F16/128 ,  G06F16/119 ,  G06F16/188 ,  G06F16/1827

Abstract: Easily provided is a file virtualization function without being affected by an application that accesses a file system.
A CPF node containerizes an application program and an IO Hook program and provides the application program and the IO Hook program to a client, the application program performs call processing on the virtual file system provided by the IO Hook program on the basis of an operation request for a file from the client, the IO Hook program performs processing for updating state management information of the file on the basis of input information or operation content with respect to a virtual file system related to the operation request, and the file virtualization program performs file management processing between a CPF and a CAS on the basis of the state management information and outputs the call processing to a distributed file system program.