Flowzilla: A Methodology for Detecting Data Transfer Anomalies in Research Networks
Flowzilla: A Methodology for Detecting Data Transfer Anomalies in Research Networks
复制标题
Flowzilla：一种检测研究网络中数据传输异常的方法
DOI：
10.1109/indis.2018.00004
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
S. Peisert
中科院分区：
文献类型：
--
作者：
Anna Giannakou;D. Gunter;S. Peisert
研究网络旨在支持跨越多个网络链接的大量科学数据传输。像任何其他网络一样，研究网络也会遇到异常现象。异常是指研究网络流量水平偏离正态分布的情况。诊断异常对于网络运营商和用户都是至关重要的（例如，科学家）。在本文中，我们提出了Flowzilla，一个通用的框架，用于检测和量化任意大小的科学数据传输的异常。Flowzilla采用随机森林回归（RFR）来预测数据传输的大小，并利用自适应阈值机制来检测离群值。我们的结果表明，我们的框架达到了高达92.5%的检测准确率。此外，我们能够预测训练后长达10周的数据传输大小，准确率超过90%。
Research networks are designed to support high volume scientific data transfers that span multiple network links. Like any other network, research networks experience anomalies. Anomalies are deviations from profiles of normality in a research network's traffic levels. Diagnosing anomalies is critical both for network operators and users (e.g., scientists). In this paper we present Flowzilla, a general framework for detecting and quantifying anomalies on scientific data transfers of arbitrary size. Flowzilla incorporates Random Forest Regression(RFR) for predicting the size of data transfers and utilizes an adaptive threshold mechanism for detecting outliers. Our results demonstrate that our framework achieves up to 92.5% detection accuracy. Furthermore, we are able to predict data transfer sizes up to 10 weeks after training with accuracy above 90%.