Thesis Details
Software-Controlled Network Traffic Monitoring
This dissertation thesis deals with the design of a novel hardware acceleration, software controlled (defined) concept for high-speed computer networks. The main goal is to propose general, flexible and easy to use acceleration platform for various network security and monitoring applications suitable for deployment in real 100\,Gbps and faster networks. The thesis starts with the survey of the current state of the art in network monitoring, security and accelerated high-speed traffic processing. Based on the survey, a brand-new concept called Software Defined Monitoring (SDM) is formulated and proposed. A key feature of the concept lies in hardware accelerated, application specific (controlled), flow based, informed reduction and distribution of captured network traffic. This brings high-speed hardware processing coupled with flexible software control, which together leads to an easy creation of various complex high-performance network applications. Further optimizations and enhancements of the main SDM concept and its selected components are also explored resulting in creation of unique and novel designs of generally usable FPGA architecture of modular packet header parser and cuckoo hash based high-throughput packet classification engine. Finally, high-speed SDM prototype using FPGA acceleration network interface card is created and thoroughly evaluated under real network conditions. Achievable performance improvements in several chosen monitoring and security use case scenarios are measured and shown. The SDM prototype is also deployed in production monitoring of real backbone network by Cesnet association and has been commercialized by Netcope Technologies.
FPGA, hardware acceleration, SDN, Software Defined Networking, monitoring, security, high-speed networks
@phdthesis{FITPT828,
author = "Luk\'{a}\v{s} Kekely",
type = "Ph.D. thesis",
title = "Software-Controlled Network Traffic Monitoring",
school = "Brno University of Technology, Faculty of Information Technology",
year = 2017,
location = "Brno, CZ",
language = "english",
url = "https://www.fit.vut.cz/study/phd-thesis/828/"
}