Adaptive Protection of Scientific Backbone Networks Using Machine Learning

Ferenc Mogyorosi,Zsolt Kenesi,Alija Pasic,Peter Revisnyei,Janos Tapolcai,Richard Cziva

doi:10.1109/tnsm.2021.3050964

Abstract

In this article, we propose a new protection scheme for backbone networks to guarantee high service availability. The presented scheme does not require any reconfiguration immediately after the failure (i.e., it is proactive). At the same time, it does not require any reserved backup network resources either. To achieve these seemingly contradictory goals, we utilize the recent advancements in Machine Learning (ML) to implement a network intelligence that periodically re-allocates the unused capacity as protection bandwidth to meet the service availability requirements of each connection. Our goal is achieved by two components (1) predicting the traffic for the next period on each link, and (2) intelligently selecting the best fit dedicated protection scheme for the next period depending on the estimated unused (spare) bandwidth and the previous service availability violations. Note that re-allocating protection bandwidth affects neither the operational connections nor the current best practice of operators to over-provision network bandwidth to support elephant flows. Finally, we provide a case study on the real traffic from Energy Sciences Network (ESnet), a high-speed, international scientific backbone network. The key benefit of our framework is that adaptively utilizing the over-provisioned bandwidth for spare capacity is sufficient to improve the availability from three-nines to five-nines (in ESnet for the 30 examined connections). The drawback is negligible bandwidth limitations; the user perceives a minor and very temporal bandwidth limitation in less than 0.1% of the time.

Highlights

I N TODAY’S connected era, communication networks are considered among the topmost critical infrastructures
This is achieved by predicting the traffic for the period on each link, and intelligently selecting the best fit dedicated protection scheme for the period depending on the estimated unused bandwidth, and the estimated probability of violating the required service availability according to the Service Level Specifications (SLSs)
We assume if the resolution of the traffic measurements is better (e.g., 100ms), there is a higher correlation between the adjacent data points; much higher precision is expected

Summary

Introduction

I N TODAY’S connected era, communication networks are considered among the topmost critical infrastructures. The new mission-critical applications such as telesurgery or stock market clearly demand a higher Quality of Service (QoS) of the underlying network infrastructure. 2) Long Short-Term Memory (LSTM) [10]: neural networks are an improvement of the recurrent neural network (RNN) [11] to avoid the vanishing or exploding gradient problem. The neural network utilizes a gated cell too, where the cell decides whether or not to store or drop the information. The purpose of these gates is to have a long term memory to learn from experiences that have a very long time lags in between. Since the LSTM neural network was built to deal with time series, it is the most used neural network for traffic prediction

Objectives

Results

Conclusion