We present the longitudinal trending analysis of traffic anomalies on a trans-Pacific backbone network over nine years. Throughout our analysis, we try to answer several questions: how frequent do such anomalies appear and how long do they last? Does a set of anomalous hosts occur correspondingly? We answer these by applying the state-of-the-art anomaly detectors to (un)anonymized packet traces and look into interesting insights from the long-term analysis. The key observations are as follow. The sources of anomalies are decreasing over the recent years, but take a significant portion of traffic volume during the measurement period (i.e., 0.03% of all IP addresses take up to 30% of traffic volume). The frequency analysis reveals that there ...
In the last years, the ever increasing number of network attacks has brought the research attention ...
The increasing practicality of large-scale flow capture makes it possible to conceive of traffic ana...
We consider the problem of traffic anomaly detection in IP networks. Traffic anomalies typically ari...
The dominating Internet protocols, IP and TCP, allow some flexibility in implementation, including a...
It is now widely accepted that packet network traffic exhibits long-range dependence (LRD), and this...
Abstract—In this paper we present methodological advances in anomaly detection, which, among other p...
In this paper we present methodological advances in anomaly detection, which, among other purposes, ...
In this paper we focus on passive measurements of TCP traffic, main component of nowadays traffic. W...
Computer networks are becoming increasingly important in supporting business and everyday activities...
Computer networks are becoming increasingly important in supporting business and everyday activities...
We use anonymized flow data collected from a 10Gbps backbone link to discover and analyze malicious ...
Dynamic networks, also called network streams, are an im-portant data representation that applies to...
The paper proposes how to detect network traffic anomalies through packet header data. In this the s...
The detection of anomalies in network traffic is an important task in today’s Internet. Among variou...
Traffic anomalies can create network congestion, so its prompt and accurate detection would allow ne...
In the last years, the ever increasing number of network attacks has brought the research attention ...
The increasing practicality of large-scale flow capture makes it possible to conceive of traffic ana...
We consider the problem of traffic anomaly detection in IP networks. Traffic anomalies typically ari...
The dominating Internet protocols, IP and TCP, allow some flexibility in implementation, including a...
It is now widely accepted that packet network traffic exhibits long-range dependence (LRD), and this...
Abstract—In this paper we present methodological advances in anomaly detection, which, among other p...
In this paper we present methodological advances in anomaly detection, which, among other purposes, ...
In this paper we focus on passive measurements of TCP traffic, main component of nowadays traffic. W...
Computer networks are becoming increasingly important in supporting business and everyday activities...
Computer networks are becoming increasingly important in supporting business and everyday activities...
We use anonymized flow data collected from a 10Gbps backbone link to discover and analyze malicious ...
Dynamic networks, also called network streams, are an im-portant data representation that applies to...
The paper proposes how to detect network traffic anomalies through packet header data. In this the s...
The detection of anomalies in network traffic is an important task in today’s Internet. Among variou...
Traffic anomalies can create network congestion, so its prompt and accurate detection would allow ne...
In the last years, the ever increasing number of network attacks has brought the research attention ...
The increasing practicality of large-scale flow capture makes it possible to conceive of traffic ana...
We consider the problem of traffic anomaly detection in IP networks. Traffic anomalies typically ari...