Network Centric Traffic Analysis

To provide more reliable and secure Internet
service, Internet service providers have more and
more interests in network centric traffic analysis.
This book considers this issue from two
perspectives, which are of ISP's most interest: 1)
network centric anomaly detection and 2) network
centric traffic classification.

In our study on network centric anomaly detection,
we designed an edge router based framework to detect
anomaly in the first place they enter network; we
proposed the so-called two-way matching features,
which are effective indicators of network anomalies;
and we creatively considered spatial and temporal
correlation among edge routers at the same time.

To tap the potential profits made by multimedia
services, ISPs are of much interest to detect voice
and video traffic. Yet, to our best knowledge no
existing approaches are available to separate
between voice and video. To solve the problem, we
creatively applied spectral analysis techniques to
extract regularities in multimedia traffic and used
minimum distance to subspace as classification
metric. Results demonstrate the effectiveness and
robustness of our approach.

Autorentext

Jieyan Fan received his Ph.D. in Electrical and Computer
Engineering from University of Florida, Gainesville, FL, in

2007. He is now working in Yahoo! Inc.

      Dapeng Wu received his Ph.D. in Electrical and Computer
      Engineering from Carnegie Mellon University, Pittsburgh, PA, in

2008. He is now an associate professor in University of Florida.

      Klappentext

      To provide more reliable and secure Internet
      service, Internet service providers have more and
      more interests in network centric traffic analysis.
      This book considers this issue from two
      perspectives, which are of ISP's most interest: 1)
      network centric anomaly detection and 2) network
      centric traffic classification.

      In our study on network centric anomaly detection,
      we designed an edge router based framework to detect
      anomaly in the first place they enter network; we
      proposed the so-called two-way matching features,
      which are effective indicators of network anomalies;
      and we creatively considered spatial and temporal
      correlation among edge routers at the same time.

      To tap the potential profits made by multimedia
      services, ISPs are of much interest to detect voice
      and video traffic. Yet, to our best knowledge no
      existing approaches are available to separate
      between voice and video. To solve the problem, we
      creatively applied spectral analysis techniques to
      extract regularities in multimedia traffic and used
      minimum distance to subspace as classification
      metric. Results demonstrate the effectiveness and
      robustness of our approach.