公开(公告)号：US11929912B2

公开(公告)日：2024-03-12

申请号：US18074319

申请日：2022-12-02

Applicant: Huawei Technologies Co., Ltd.

Inventor： Mahmoud Mohamed Bahnasy ,  Fenglin Li ,  Shihan Xiao ,  Xiangle Cheng ,  Majid Ghaderi

IPC: H04L45/02 ,  H04L45/00

CPC classification number: H04L45/08 ,  H04L45/14

Abstract: A method, processing system and processor-readable medium for generating network configurations using a graph neural network (GNN) are provided. The method may include receiving a first matrix M generated based on a set of network requirements; storing a GNN having a plurality of nodes v and a plurality of edges; initializing the GNN based on a second matrix Xv having a plurality of elements, each element corresponding to a node from a plurality of nodes v of the GNN; and generating an output matrix having a plurality of nodes labelled based on the first matrix M.

公开(公告)号：US20240022512A1

公开(公告)日：2024-01-18

申请号：US18471725

申请日：2023-09-21

Applicant: Huawei Technologies Co., Ltd.

Inventor： Shihan Xiao ,  Bo Wu ,  Haibo Wang ,  Xiaofei Xu

IPC: H04L45/745 ,  G06N3/08

CPC classification number: H04L45/745 ,  G06N3/08 ,  H04L49/55

Abstract: In a packet processing method, a network device obtains a first packet. The network device determines a first index value based on address information of the first packet and a neural network model. The network device determines, from an action information table based on the first index value, action information corresponding to the first packet, where the action information table includes at least one entry, and each entry corresponds to an index value and a piece of action information. The network device processes the first packet based on the action information corresponding to the first packet.