During my Thesis I was engaged in experimenting with hybrid SDN, which is a part of the evolution of Software Defined Networking. This topic is a really important turning point in the evolution of networking, more and more researches are made which open up new part of this topic and the main providers are also more and more involved because they will always want to upgrade their networking capabilities, thus improve their performance with the actual infrastucture or with using newer technologies.
Examining the current researches, it’s still an open question what methods and tools are the best way to achieve a cost-effective result. It was my duty to look into some of these researches and to get more knowledge on the oppurtunities offered by this SDN paradigm and to try out some of them on a network that is as close to a physically connected network as possible. In a scenario like that it is not trivial how we can get our networking devices to communicate with each other, what protocolos we should use, or how can we test these layouts. During establishing the goals of this project, I found tasks which covered broadly the topic, so beside making a working network I had to make steps toward implementing a brand new solution. The main goal of this project was to create a network as close to a physically connected network as possible, so I could not use a method like virtualize everything on a single host.
In the first part of my Thesis I present some cases of hybrid networks where the hybrid word means thath hybrid nodes are used, and then I continue with a case where our SDN devices communicate with so-called legacy devices. The majority of my work consisted of creating the planned network step-by-step and to make it less virtualized. In the last section I present a specification for a system in which legacy devices and SDN devices can use routing protocols together and so share information about the whole network.