The appearance of the software defined networks and cloud computing generates new network virtualization challenges, and generates new expectations over against the service providers. It is important to investigate the virtual network embedding in this environment, because the already existing algorithms (such as DViNE-PS  and MaxMatch-PS ) working with good results, but only in static environments. But in the reality the request arrives continually, or with other words, dynamically to the service providers. In my thesis I examined the VNE problem and studied five of the static algorithms. These gave me a good base to a dynamic algorithm.
In my thesis I write about the ALEVIN simulation framework I had chosen. I write about the benefits and the limits of the framework. I evaluated the five formerly studied algorithms, until the point, where I was able to choose the best of them based on the average running time or the average revenue of the algorithms. Based on the tests I had chosen the MaxMatch-PS algorithm for demonstration of how an algorithm works in dynamic environment. To do this, I had to extend it with two crucial elements. With this two element or aspects in mind, every other algorithm could have been improved, but I stayed with the one, which had the best results during the tests.
The ALEVIN simulation framework was not prepared for simulating dynamic environments, so I had made improvements on the framework too. One of these key improvements was adding a new functionality, so I was able to add new randomly generated virtual networks so, that the older physical and virtual networks was reserved in their former state. Additionally I created a new form of the MaxMatch-PS algorithm, so I was able to evaluate the dynamic and the static version under the same conditions. All the results look like, what I was expecting: the dynamic version of the algorithm were slower, because of the plus work it had to make, but it had a lot better performance in mapping the VNE request to the physical network, then the static one.