公开(公告)号：US20230275848A1

公开(公告)日：2023-08-31

申请号：US18311430

申请日：2023-05-03

Applicant: Hewlett Packard Enterprise Development LP

Inventor： Ali Tariq ,  Lianjie Cao ,  Faraz Ahmed ,  Puneet Sharma

IPC: H04L47/80 ,  H04L47/78 ,  H04L43/16 ,  H04L47/70 ,  H04L43/0882 ,  H04L47/762

CPC classification number: H04L47/803 ,  H04L47/781 ,  H04L43/16 ,  H04L47/822 ,  H04L43/0882 ,  H04L47/762

Abstract: Systems and methods are provided for updating resource allocation in a distributed network. For example, the method may comprise allocating a plurality of resource containers in a distributed network in accordance with a first distributed resource configuration. Upon determining that a processing workload value exceeds a stabilization threshold of the distributed network, determining a resource efficiency value of the plurality of resource containers in the distributed network. When a resource efficiency value is greater than or equal to the threshold resource efficiency value, the method may generate a second distributed resource configuration that includes a resource upscaling process, or when the resource efficiency value is less than the threshold resource efficiency value, the method may generate the second distributed resource configuration that includes a resource outscaling process. The resource allocation may transmit the second to update the resource allocation.

公开(公告)号：US20240289421A1

公开(公告)日：2024-08-29

申请号：US18654953

申请日：2024-05-03

Applicant: HEWLETT PACKARD ENTERPRISE DEVELOPMENT LP

Inventor： Lianjie Cao ,  Faraz Ahmed ,  Puneet Sharma ,  Ali Tariq

IPC: G06F18/214 ,  G06F9/50 ,  G06F11/34 ,  G06F18/2415 ,  G06N20/00

CPC classification number: G06F18/214 ,  G06F9/5022 ,  G06F9/5027 ,  G06F9/505 ,  G06F9/5061 ,  G06F11/3414 ,  G06F18/24155 ,  G06N20/00

Abstract: Systems and methods can be configured to determine a plurality of computing resource configurations used to perform machine learning model training jobs. A computing resource configuration can comprise: a first tuple including numbers of worker nodes and parameter server nodes, and a second tuple including resource allocations for the worker nodes and parameter server nodes. At least one machine learning training job can be executed using a first computing resource configuration having a first set of values associated with the first tuple. During the executing the machine learning training job: resource usage of the worker nodes and parameter server nodes caused by a second set of values associated with the second tuple can be monitored, and whether to adjust the second set of values can be determined. Whether a stopping criterion is satisfied can be determined. One of the plurality of computing resource configurations can be selected.

公开(公告)号：US12132668B2

公开(公告)日：2024-10-29

申请号：US18311430

申请日：2023-05-03

Applicant: Hewlett Packard Enterprise Development LP

Inventor： Ali Tariq ,  Lianjie Cao ,  Faraz Ahmed ,  Puneet Sharma

IPC: H04L47/78 ,  H04L43/0882 ,  H04L43/16 ,  H04L47/70 ,  H04L47/762 ,  H04L47/80

CPC classification number: H04L47/803 ,  H04L43/0882 ,  H04L43/16 ,  H04L47/762 ,  H04L47/781 ,  H04L47/822

Abstract: Systems and methods are provided for updating resource allocation in a distributed network. For example, the method may comprise allocating a plurality of resource containers in a distributed network in accordance with a first distributed resource configuration. Upon determining that a processing workload value exceeds a stabilization threshold of the distributed network, determining a resource efficiency value of the plurality of resource containers in the distributed network. When a resource efficiency value is greater than or equal to the threshold resource efficiency value, the method may generate a second distributed resource configuration that includes a resource upscaling process, or when the resource efficiency value is less than the threshold resource efficiency value, the method may generate the second distributed resource configuration that includes a resource outscaling process. The resource allocation may transmit the second to update the resource allocation.

公开(公告)号：US12001511B2

公开(公告)日：2024-06-04

申请号：US17199294

申请日：2021-03-11

Applicant: HEWLETT PACKARD ENTERPRISE DEVELOPMENT LP

Inventor： Lianjie Cao ,  Faraz Ahmed ,  Puneet Sharma ,  Ali Tariq

IPC: G06F18/214 ,  G06F9/50 ,  G06F11/30 ,  G06F11/34 ,  G06F18/2415 ,  G06N3/0464 ,  G06N3/063 ,  G06N3/0985 ,  G06N7/01 ,  G06N20/00 ,  G06V40/16

CPC classification number: G06F18/214 ,  G06F9/5022 ,  G06F9/5027 ,  G06F9/505 ,  G06F9/5061 ,  G06F11/3414 ,  G06F18/24155 ,  G06N20/00

Abstract: Systems and methods can be configured to determine a plurality of computing resource configurations used to perform machine learning model training jobs. A computing resource configuration can comprise: a first tuple including numbers of worker nodes and parameter server nodes, and a second tuple including resource allocations for the worker nodes and parameter server nodes. At least one machine learning training job can be executed using a first computing resource configuration having a first set of values associated with the first tuple. During the executing the machine learning training job: resource usage of the worker nodes and parameter server nodes caused by a second set of values associated with the second tuple can be monitored, and whether to adjust the second set of values can be determined. Whether a stopping criterion is satisfied can be determined. One of the plurality of computing resource configurations can be selected.

公开(公告)号：US11665106B2

公开(公告)日：2023-05-30

申请号：US17468517

申请日：2021-09-07

Applicant: HEWLETT PACKARD ENTERPRISE DEVELOPMENT LP

Inventor： Ali Tariq ,  Lianjie Cao ,  Faraz Ahmed ,  Puneet Sharma

IPC: H04L43/16 ,  H04L43/0882 ,  H04L47/80 ,  H04L47/78 ,  H04L47/70 ,  H04L47/762

CPC classification number: H04L47/803 ,  H04L43/0882 ,  H04L43/16 ,  H04L47/762 ,  H04L47/781 ,  H04L47/822

Abstract: Systems and methods are provided for updating resource allocation in a distributed network. For example, the method may comprise allocating a plurality of resource containers in a distributed network in accordance with a first distributed resource configuration. Upon determining that a processing workload value exceeds a stabilization threshold of the distributed network, determining a resource efficiency value of the plurality of resource containers in the distributed network. When a resource efficiency value is greater than or equal to the threshold resource efficiency value, the method may generate a second distributed resource configuration that includes a resource upscaling process, or when the resource efficiency value is less than the threshold resource efficiency value, the method may generate the second distributed resource configuration that includes a resource outscaling process. The resource allocation may transmit the second to update the resource allocation.