Activity Number:
|
15
- Networks, Multivariate Analysis, and Time Series
|
Type:
|
Topic Contributed
|
Date/Time:
|
Sunday, July 30, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Royal Statistical Society
|
Abstract #322988
|
|
Title:
|
A Framework for the Online Monitoring and Analysis of Multi-Gigabit Network Streams
|
Author(s):
|
Stilian Stoev* and Michalis Kallitsis and George Michailidis and Shrijita Bhattacharya and Zheng Gao
|
Companies:
|
University of Michigan and Merit Network and University of Florida and University of Michigan and University of Michigan
|
Keywords:
|
hashing ;
data stream ;
extremes ;
heavy tails ;
network monitoring ;
online algorithms
|
Abstract:
|
We shall describe a recently developed platform for collecting meta-data on multi-10gigabit network streams. The AMON (All Packet MONitoring) platform utilizes efficient hashing techniques to build statistical summaries of high-volume and high-velocity data streams observed by wide-area network Internet service providers. Online statistical algorithms are developed for the detection of changes in network structure and connectivity associated with cyber-attacks such as distributed denial of service (DDoS) attacks. The methodology is illustrated with real data from Merit Network -- a large Internet service provider in Michigan. We sketch some future challenges on the probabilistic modeling and statistical inference for such types of data.
|
Authors who are presenting talks have a * after their name.