GPON ATTACKS AND ERRORS CLASSIFICATION

GPON ATTACKS AND ERRORS CLASSIFICATION

Paper in proceedings outside WoS and Scopus

This paper focuses on various types of attacks and errors in an activation process of Gigabit-capable passive optical networks. The process sends messages via Physical Layer Operation Administration and Maintenance header field inside the transmitted frame. An exemplar network communication is captured by a special hardware-accelerated network interface card capable of processing optical signals from passive optical networks. The captured data is filtered of irrelevant parts and messages and correctly formatted into a suitable shape for a neural network. The filtered data is divided into small sequences called time windows and analyzed using a recurrent neural network-based on Gated recurrent unit cells. A new neural network model is designed to classify sequences into several categories: additional message, missing message, error inside (noisy) message, and message order error. All of these categories represent a certain type of attack or error. The proposed model can distinguish message sequences into these categories with high accuracy resulting in revealing a possible attacker or drift from protocol recommendation.

This paper focuses on various types of attacks and errors in an activation process of Gigabit-capable passive optical networks. The process sends messages via Physical Layer Operation Administration and Maintenance header field inside the transmitted frame. An exemplar network communication is captured by a special hardware-accelerated network interface card capable of processing optical signals from passive optical networks. The captured data is filtered of irrelevant parts and messages and correctly formatted into a suitable shape for a neural network. The filtered data is divided into small sequences called time windows and analyzed using a recurrent neural network-based on Gated recurrent unit cells. A new neural network model is designed to classify sequences into several categories: additional message, missing message, error inside (noisy) message, and message order error. All of these categories represent a certain type of attack or error. The proposed model can distinguish message sequences into these categories with high accuracy resulting in revealing a possible attacker or drift from protocol recommendation.

Activation Process, GPON, GRU, Recurrent Neural Network, PLOAM

Activation Process, GPON, GRU, Recurrent Neural Network, PLOAM

TOMAŠOV, A.

2021

13.07.2021

Brno University of Technology, Faculty of Electrical Engineering and Communication

Brno

332

336

5

https://www.fekt.vut.cz/conf/EEICT/archiv/sborniky/EEICT_2021_sbornik_1.pdf