| "Datacenter networks and wide area networks are
two major network fields facing constantly changing
challenges due to the quick adoption of new network
technologies. Because of the large volume of traffic
involved in these two scenarios, the successful detection
of abnormal or malicious network activities relies
heavily on an effective first-line screening to reduce the
traffic as low in volume and relevant to the detection as
possible. While the data reduction can be achieved in
many ways, a right threshold is the key to keep the
screening effective and accurate. This paper analyzed the
characteristics and requirements of the irregularity
detection of both datacenter and backbone networks,
and proposed a method adaptively determine the
threshold for the screening and detection system."
|
