Unlock instant, AI-driven research and patent intelligence for your innovation.

Cluster link detection method and system based on deep learning

A link detection and deep learning technology, applied in the field of cluster link detection based on deep learning, can solve the problems of link packet loss, slow data transmission, and inability to detect link detection methods.

Active Publication Date: 2020-12-01
INSPUR SUZHOU INTELLIGENT TECH CO LTD
View PDF5 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, when the cluster system processes computing requests, it often encounters the problem of link packet loss or slow data transmission, so link status detection is particularly important
There is randomness in the link problem. In the existing technology, the link detection method cannot be detected by the existing general means, and can only be based on the empirical knowledge of the specific cluster system as the basis for detection.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Cluster link detection method and system based on deep learning
  • Cluster link detection method and system based on deep learning
  • Cluster link detection method and system based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0025] Embodiment 1 of the present invention proposes a cluster link detection method based on deep learning, the combination of deep learning and cluster system, using the empirical learning of deep learning method (such as convolutional neural network) to generate the link detection of the cluster system model, and provide judgment basis for link optimization and path aggregation.

[0026] Such as figure 1 It is a schematic diagram of a link detection model in a deep learning-based cluster link detection method proposed in Embodiment 1 of the present invention.

[0027] First obtain the state value of the cluster link node, use the empirical learning of the deep learning method (such as convolutional neural network) to perform model training on the massive data of the state value of the cluster link node, and stop the training after the model converges. Then, according to the converged model, the state of the cluster link nodes is arbitrated and aggregated through the route...

Embodiment 2

[0044] Based on the present invention, a deep learning-based cluster link detection method is proposed, and a deep learning-based cluster link detection system is also proposed, such as image 3 It is a schematic diagram of a cluster link detection system based on deep learning proposed in Embodiment 2 of the present invention.

[0045] The system includes: a training module and a judging module; the training module is used to use the cluster link state value as the input feature value of the convolutional neural network; the convolutional neural network trains the massive data of the cluster link state value, and converges in the training model After that, the training is stopped; the judging module is used to retain the links that meet the threshold according to the model after convergence, and perform link merging on the links that do not meet the threshold.

[0046] The process performed by the training module is: quantitatively collect the cluster link status value and sa...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a cluster link detection method and system based on deep learning. The method comprises the following steps: taking a cluster link state value as an input characteristic value of a convolutional neural network; training the mass data of the cluster link state value by the convolutional neural network, and stopping training after convergence of the training model; and according to the converged model, carrying out link reservation on the links conforming to the threshold value, and carrying out link combination on the links not conforming to the threshold value. The cluster link state value comprises link bandwidth, link throughput and link transmission delay. Based on the cluster link detection method provided by the invention, the invention also provides a cluster link detection system based on deep learning. According to the method, the automatic tuning characteristic of the convolutional neural network is fully utilized, and specific processing is carried outon the characteristics of the impassable clusters. Through the detection model obtained by training, each node resource in the cluster system can be utilized to the greatest extent, so that the automation and computing performance of the cluster system are improved.

Description

technical field [0001] The invention belongs to the technical field of distributed cluster detection, and in particular relates to a deep learning-based cluster link detection method and system. Background technique [0002] Distributed clusters have become the key technology of today's high-performance computing. In many fields, distributed clusters have played a very important role. By building high-performance clusters, it can solve the problem of small calculation amount and long operation cycle of a single computer. [0003] However, when the cluster system processes computing requests, it often encounters the problem of link packet loss or slow data transmission, so link status detection is particularly important. There is randomness in the link problem. In the prior art, the link detection method cannot be detected by the existing general means, and the empirical knowledge of the specific cluster system can only be used as the basis for detection. Contents of the in...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): H04L12/24G06N3/04G06N3/08
CPCH04L41/145H04L41/0823H04L41/042G06N3/08G06N3/045Y02D10/00
Inventor 葛晨
Owner INSPUR SUZHOU INTELLIGENT TECH CO LTD