Scribd
Upload
35 views

Title: Sruthi

This paper proposes a neural network model called network-in-network (NIN) for classifying encrypted network traffic. NIN uses convolutional layers with micro networks to better model local features. It also uses global average pooling instead of fully connected layers to reduce parameters. The paper builds deep NIN models and designs a parallel strategy to process packet headers and bodies separately to improve classification accuracy. Experiments on an encrypted traffic dataset show NIN achieves better accuracy than CNNs, and the parallel strategy further improves performance of single NIN models for traffic classification.

Uploaded by

Praveen Kumar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
35 views

Title: Sruthi

This paper proposes a neural network model called network-in-network (NIN) for classifying encrypted network traffic. NIN uses convolutional layers with micro networks to better model local features. It also uses global average pooling instead of fully connected layers to reduce parameters. The paper builds deep NIN models and designs a parallel strategy to process packet headers and bodies separately to improve classification accuracy. Experiments on an encrypted traffic dataset show NIN achieves better accuracy than CNNs, and the parallel strategy further improves performance of single NIN models for traffic classification.

Uploaded by

Praveen Kumar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

TITLE

SRUTHI
ABSTRACT
• Network traffic classification aims to recognize different application or traffic types by

analyzing received data packets. This paper presents a neural network model with deep and parallel

network-in-network (NIN) structures for classifying encrypted network traffic.

• Comparing with standard convolutional neural networks (CNN), NIN adopts a micro network

after each convolution layer to enhance local modeling.

• Besides, NIN utilizes a global average pooling instead of traditional fully connected layers

before final classification, which reduces the number of model parameters significantly.

• In our proposed method, deep NIN models with multiple MLP convolutional layers are built to

map fixed-length packet vectors towards application or traffic labels.

• Furthermore, a parallel decision strategy of building two subnetworks to process packet header

and packet body separately is designed considering that they may carry different kinds of clues

for classification.

• The results of our experiments on the “ISCX VPN-nonVPN” encrypted traffic dataset show

that NIN models can achieve a better balance between classification accuracy and model

complexity than conventional CNNs.


EXISTING SYSTEM
• Since 2006, deep learning has emerged as a new area of machine learning research .

• Deep learning models, uch as deep neural networks (DNNs), convolutional neural

networks (CNNs) and recurrent neural networks (RNNs), have be applied to various research

areas, e.g., image classification, speech recognition , and natural language

processing, and have achieved significant progresses.

• Comparing with conventional statistical classifiers, deep learning models are better at

describing the complex and nonlinear mapping relationship from input features towards

class labels.

• Besides, deep learning models are able to learn feature representations automatically from raw

data, which alleviates the dependency on manually designed features and simplifies the

pipeline of building classifiers. Therefore, such deep learning models have been introduced

into machine learning based encrypted traffic classification recently.


PROPOSED SYSTEM
• This project presents a neural network model with deep and parallel network-in-network (NIN) structures for encrypted
traffic classification.
• The NIN model was initially proposed for image classification . Comparing with CNN, NIN adopts a micro network after
each convolution layer to enhance its local modeling and abstraction ability.
• Besides, NIN utilizes a global average pooling instead of traditional fully connected layers before final classification,which
reduces the number of model parameters significantly.
• In this paper, deep NIN models with multiple convolutional layers are built for encrypted traffic classification, in which the
micro network after each convolutional layer is an MLP.
• Furthermore, a parallel decision strategy is designed, which trains two sub-networks to process packet headers and packet
bodies separately.
• The final classification result for a data packet is determined by fusing the decisions of both subnetworks.
• We evaluated the performance of our proposed methods on the “ISCX VPN-nonVPN” encrypted traffic dataset.
• Experimental results show that NIN models can achieve a better balance between classification accuracy and model
complexity than conventional CNNs.
• The parallel decision strategy can further improve the accuracy of single NIN models for encrypted network traffic
classification.
• The best test set F1 scores of traffic characterization and application identification are 0.983 and 0.985 respectively.

•  
Hardware Requirements: 

• System : Pentium IV 2.4 GHz.


• Hard Disk : 40 GB.
• Floppy Drive : 1.44 Mb.
• Monitor : 15 VGA Colour.
• Mouse : Logitech.
• Ram : 256 Mb
Software Requirements:

• Operating system :- Windows XP Professional


• Front End :JAVA, Swing(JFC),RMIJ2ME
• Back End : MY SQL
• Tool :NETBEANS
CONCLUSION

• This project has presented a method of building deep and parallel network-in-network (NIN)

models for encrypted network traffic classification. This method aims at mapping fixed-length

data packets towards the labels of application or traffic categories. Based on deep NIN networks

with multiple MLP convolutional modules, a parallel decision strategy is designed which builds

two sub-networks for processing packet header and packet body separately. Experimental results

on the “ISCX VPN-nonVPN” encrypted traffic dataset show that NIN models achieved better

performance than CNNs. Besides, the parallel decision strategy further improved the accuracy of

using single NIN model for traffic classification. To boost the performance of encrypted traffic

classification using only Application Layer data will be a task of our future work.

 
REFERENCES

• [1] A. Dainotti, A. Pescape, and K. C. Claffy, “Issues and future directions in traffic classification,” IEEE network, vol. 26, no.

1, pp. 35–40, 2012.

• [2] A. W. Moore and K. Papagiannaki, “Toward the accurate identification of network applications,” in International

Workshop on Passive and Active Network Measurement. Springer, 2005, pp. 41–54.

• [3] M. Finsterbusch, C. Richter, E. Rocha, J.-A. Muller, and K. Hanssgen, “A survey of payload-based traffic classification

approaches,” IEEE Communications Surveys & Tutorials, vol. 16, no. 2, pp. 1135–1156, 2013.

• [4] B. Yamansavascilar, M. A. Guvensan, A. G. Yavuz, and M. E. Karsligil, “Application identification via network traffic

classification,” in 2017 International Conference on Computing, Networking and Communications (ICNC). IEEE, 2017, pp.

843–848.

• [5] G. Draper-Gil, A. H. Lashkari, M. S. I. Mamun, and A. A. Ghorbani, “Characterization of encrypted and VPN traffic

using time-related,” in Proceedings of the 2nd international conference on information systems security and privacy

(ICISSP), 2016, pp. 407–414.

• [6] R. Yuan, Z. Li, X. Guan, and L. Xu, “An svm-based machine learning method for accurate internet traffic classification,”

Information Systems Frontiers, vol. 12, no. 2, pp. 149–156, 2010.

• [7] G. Hinton, S. Osindero, and Y.-W. Teh, “A fast learning algorithm for deep belief nets,” Neural Computation, vol. 18, no.

7, pp. 1527–1554, 2006

You might also like