公开(公告)号：US20240036838A1

公开(公告)日：2024-02-01

申请号：US17816031

申请日：2022-07-29

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Jun Su ,  Peng Hui Jiang ,  Gang Tang ,  Guang Han Sui ,  ZI YUN KANG

IPC: G06F8/41 ,  H04L9/40 ,  G06F9/455

CPC classification number: G06F8/433 ,  H04L63/0281 ,  G06F9/45558 ,  G06F2009/45595 ,  G06F2009/4557

Abstract: This disclosure provides a method, a computing system and a computer program product for managing resource sharing among devices with different platforms. The method comprises obtaining connection information for connecting to a requesting device by decoding authentication information, wherein the authentication information is created based at least on an identifier of the requesting device. The method further comprises sending a request including the connection information to a request proxy, wherein the request is redirected by the request proxy to a plurality of devices having more than one platform. The method further comprises receiving a response to the request, wherein the response is obtained by combining responses received by the request proxy from the plurality of devices.

公开(公告)号：US20230367869A1

公开(公告)日：2023-11-16

申请号：US17662710

申请日：2022-05-10

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： NAIJIE LI ,  Min Cheng ,  Kui Zhang ,  Yi Chai ,  Guang Han Sui

IPC: G06F21/51 ,  G06F9/54 ,  G06F12/14

CPC classification number: G06F21/51 ,  G06F9/545 ,  G06F12/14

Abstract: Embodiments of the present disclosure relate to a method, system and computer program product for providing system services. In some embodiments, a method is disclosed. According to the method, from a user program in a user address space, a request for a system service is received via a program call instruction of a set of program call instructions in an application interface code library. Based on the program call instruction, a target authorized address space of a plurality of authorized address spaces and a target system service routine for providing the system service in the target authorized address space is determined. A result of running the target system service routine in the target authorized address space is returned to the user program as a response to the request.

公开(公告)号：US11799945B2

公开(公告)日：2023-10-24

申请号：US17376297

申请日：2021-07-15

Applicant: International Business Machines Corporation

Inventor： Guang Han Sui ,  Wei Ge ,  Juan Yang ,  Lan Zhe Liu ,  Le Yao ,  Li Jun BJ Zhu

IPC: G06F15/16 ,  H04L67/06 ,  G06F16/182

CPC classification number: H04L67/06 ,  G06F16/183

Abstract: A block of data intended for a set of receiving computer systems comprising a first system and a second system is divided into a set of equal-size portions. A first portion of the set of portions is transmitted from a first file server storing the block of data to the first system. The first portion is relayed from the first file server to a second file server concurrently with the transmitting. The first portion of the set of portions is transmitted from the second file server to the second system.

公开(公告)号：US20230289214A1

公开(公告)日：2023-09-14

申请号：US17688958

申请日：2022-03-08

Applicant: International Business Machines Corporation

Inventor： Gang Tang ,  Peng Hui Jiang ,  Guang Han Sui ,  Liang Rong ,  Jun Su

IPC: G06F9/48 ,  G06F9/50 ,  G06F9/54

CPC classification number: G06F9/4881 ,  G06F9/5038 ,  G06F9/546

Abstract: An approach for intelligent task message queue management may be provided herein. The approach may include receiving a task message from a producer at an intelligent task message queue management service. The approach may include assigning a subclass to the task message based on a predetermined set of rules. The approach may include assigning the task message to a one of a plurality of subclass worklist queues based on the assigned subclass. The approach may include generating a subclass score for the task message based on a weight associated with the assigned subclass work queue. The approach may include dispatching the task message to a consumer based on the generated subclass score for the task message.

公开(公告)号：US20230267005A1

公开(公告)日：2023-08-24

申请号：US17651943

申请日：2022-02-22

Applicant: International Business Machines Corporation

Inventor： NAIJIE LI ,  Bao Zhang ,  Ming Ran Liu ,  Yuan Zhai ,  Jia Yu ,  Peng Hui Jiang ,  Guang Han Sui

IPC: G06F9/50 ,  G06F9/48 ,  G06F9/455

CPC classification number: G06F9/5022 ,  G06F9/4881 ,  G06F9/5055 ,  G06F9/45504

Abstract: A thread management process where a storage area template is generated for a plurality of threads that are to be created by a process in a computing system, the storage area template comprising at least one module template for at least one library that is to be loaded by the process. In the computing system, a plurality of storage areas is allocated for the plurality of threads based on the storage area template, respectively. The plurality of threads are managed based on the plurality of storage areas, respectively. All threads and libraries may be managed in a uniform way according to the storage area template. Therefore, thread management may be simplified and then may lead to reduced costs in the storage and time resources.

公开(公告)号：US20230236946A1

公开(公告)日：2023-07-27

申请号：US17585695

申请日：2022-01-27

Applicant: International Business Machines Corporation

Inventor： Yan Li ,  Run Qian Bj Chen ,  Chen Guang Zhao ,  Qin Qin Zhou ,  Guang Han Sui ,  Jing Li ,  You Bing Li ,  Yu Xiang Chen

IPC: G06F11/34 ,  G06F9/50

CPC classification number: G06F11/3442 ,  G06F9/5077 ,  G06F9/505

Abstract: Solutions preparing container images and data for container workloads prior to start times of workloads predicted through workload trend analysis. Local storage space on the node is managed based on workload trends, optimizing local storage of image files without requiring frequent reloading and/or deletion of image files, avoiding network intensive I/O operations when pulling images to local storage by workload scheduling systems. Systems perform collection of historical data including image and workload properties; analyze historical data for workload trends, including predicted start times, image files needed, number of nodes and types of nodes. Based on predicted future workload start times, nodes are selected from an ordered list of node requirements and workload properties. Selected nodes' local storage is managed using predicted future start times of workloads, to avoid removing image files having sooner start times, while removing (as needed) images files predictively utilized for workloads further into the future.

公开(公告)号：US11573831B2

公开(公告)日：2023-02-07

申请号：US15627660

申请日：2017-06-20

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Xiao Jie Li ,  Zhimin Lin ,  Jinming Lv ,  Guang Han Sui ,  Hao Zhou

IPC: G06F9/50 ,  G06F9/48

Abstract: Embodiments for optimizing resource usage in a distributed computing environment. Resource usage of each task in a set of running tasks associated with a job is monitored to collect resource usage information corresponding to each respective task. A resource unit size of at least one resource allocated to respective tasks in the set of running tasks is adjusted based on the resource usage information to improve overall resource usage in the distributed computing environment.

公开(公告)号：US20230031636A1

公开(公告)日：2023-02-02

申请号：US17387125

申请日：2021-07-28

Applicant: International Business Machines Corporation

Inventor： Lin Dong ,  Dong Xie ,  Jing Li ,  Guang Han Sui ,  Xiao Tian Xu

IPC: G06N5/04 ,  G06F11/34 ,  G06F16/901

Abstract: Aspects of the invention include systems and methods configured to provide simplified and efficient artificial intelligence (AI) model deployment. A non-limiting example computer-implemented method includes receiving an AI model deployment input having pre-process code, inference model code, and post-process code. The pre-process code is converted to a pre-process graph. The inference model and the post-process model are similarly converted to an inference graph and a post-process graph, respectively. A pipeline path is generated by connecting nodes in the pre-process graph, the inference graph, and the post-process graph. The pipeline path is deployed as a service for inference.

公开(公告)号：US20220398250A1

公开(公告)日：2022-12-15

申请号：US17343878

申请日：2021-06-10

Applicant: International Business Machines Corporation

Inventor： Guang Han Sui ,  Peng Hui Jiang ,  Jia Tian Zhong ,  Jun Su

IPC: G06F16/2455 ,  G06F16/28 ,  G06F11/34 ,  G06F11/20 ,  G06F11/14

Abstract: An embodiment includes deriving usage data associated with records of a database by monitoring requests to perform read operations on the records of the database. The embodiment generates record correlation data representative of correlations between respective groups of records of the database by parsing the usage data associated with the records of the database. The embodiment stores a plurality of records received as respective write requests during a first time interval in an intermediate storage medium. The embodiment identifies a correlation in the record correlation data between a first record of the plurality of records and a second record of the plurality of records. The embodiment selects, responsive to identifying the correlation, a first location in the database for writing the first record and a second location in the database for writing the second record based on a proximity of the first location to the second location.

公开(公告)号：US11526379B2

公开(公告)日：2022-12-13

申请号：US16446180

申请日：2019-06-19

Applicant: International Business Machines Corporation

Inventor： Ping Xiao ,  Peng Hui Jiang ,  Xin Peng Liu ,  Guang Han Sui

IPC: G06F9/46 ,  G06F9/50 ,  G06F9/445 ,  G06F8/20 ,  G06F8/30

Abstract: Embodiments of the present disclosure relate to a method for building an application. According to the method, a request is received from a building environment to acquire at least one component for executing at least one function of at least one feature of the application. The at least one feature is to be deployed to at least one target node in a distributed service platform comprising a plurality of nodes. The at least one target node and the at least one component are determined based on the request. The at least one component is acquired from the at least one target node. The at least one component is sent to the building environment for building the at least one feature.