Nowadays to maintain high-speed computer networks, it is an absolute necessity to visualize traffic matrices, enabling you to be well-acquainted with complex traffic patterns of the network.
Maintaining high-speed computer networks is never without problems. Small hiccups in a core network could cause trouble for thousands of people if not millions. In our technological age it can disrupt businesses regardless of which side of the network you are on, operator or end user. This is the main reason why anybody exerting control over such networks should be very well equipped and be ready when any trouble arises to combat is with utmost expertise to keep it up and running smoothly. For this they need tools that are able to help them to see the roots of problems and get rid of them promptly and effectively. One of those tools is the so called traffic matrix. This is a flexible, effective and easy to use aid in understanding the current traffic situation in the network. Unfortunately due to the excessive resources required to calculate them, they are rarely used.
My primary aim in this thesis is to design and create a system capable of effectively calculating and visualizing traffic matrices. To be able to do that and understand its basic components, knowledge of network monitoring tools is needed, hence my fist topic.
To calculate a traffic matrix, first traffic data from the network is needed. The only way for successfully obtaining that from a high-speed core network is to be equipped with a special, purpose-built measurement system. For this, the BME TMIT co-developed SCALOPES C-board proved to be an excellent choice, mainly due to its easy accessibility and availability during my work. The output of this measurement board is the input for my traffic matrix calculator and presenter system.
Visualization tools, capable of visualizing traffic matrices have been in the interest of researchers in the past few years. After reviewing a few of their proposed solutions I will introduce my own, which focuses on the need of the core network operators as opposed to being handy tools in discovering intrusions into some protected networks.
By combining existing measurement and newly created processing and presentation tools, I have engineered a system that is capable of providing useful information for network operators. It is able to show multiple, time-dependent internal (VPN) and external traffic parameters of a portion of a network (may be a selected user), all in a flexible, effective and usable way, and which were the original requirements for the newly created system.