A Fault Correlation Approach to Detect Performance Anomalies in Virtual Network Function Chains

Abstract

Network Function Virtualization is an emerging paradigm to allow the creation, at software level, of complex network services by composing simpler ones. However, this paradigm shift exposes network services to faults and bottlenecks in the complex software virtualization infrastructure they rely on. Thus, NFV services require effective anomaly detection systems to detect the occurrence of network problems. The paper proposes a novel approach to ease the adoption of anomaly detection in production NFV services, by avoiding the need to train a model or to calibrate a threshold. The approach infers the service health status by collecting metrics from multiple elements in the NFV service chain, and by analyzing their (lack of) correlation over the time. We validate this approach on an NFV-oriented Interactive Multimedia System, to detect problems affecting the quality of service, such as the overload, component crashes, avalanche restarts and physical resource contention

LSTM-based radiography for anomaly detection in softwarized infrastructures

“…Authors in [23] propose an unsupervised learning approach based on correlation variation algorithm to predict performance anomalies in VNF service chaining. It infers the service health status by collecting metrics from multiple elements in the VNF chain analyzing their correlation across time.…”

“…AEs can be used to detect anomalies as the decoder block compresses the data input dimensionality. Assuming that input data has certain correlation level [23], it can be embedded into a lower dimensional subspace, where anomalous samples are perceived significantly different which makes the reconstruction error increase significantly. AEs are considered as auto-supervised NN, as the target value is the input itself, so no labels are required in the training phase: for our targeted virtualized network infrastructure environment, this factor streamlines learning as labelling anomalies is at least difficult, if not sometimes impossible, due to the great extent of faults and threats that can affect NFV environments.…”

Machine Learning Approaches to Early Fault Detection and Identification in NFV Architectures

“…Pham et al [5] use a failure injection approach to construct a database with failure profiles, then with a Top-K Nearest Faults query algorithm they recognise a specific fault pattern and pinpoint the root cause. Authors of [6] use a statistical fault diagnosis, with Pearson correlation computed pairwise between neighboring nodes to detect anomalies. It is based on the fact that the Virtualized Network Functions (VNF) chain can be seen as a pipeline, where the input of the next network function of the chain is the output of the one before.…”

GeminiNet: Combine Fully Convolution Network With Structure of Receptive Fields for Object Detection

“…With the rapid development of information technology, deep learning-based detection technology is increasingly becoming an important component of our daily activities, such as face detection [2], fault detection [3], and disease…”