A stochastic minizinc solution to solve a software defined network configuration problem

In this project we conisider a simulated SDN net (Software Defined Network). This network is composed by a certain number of domains, and into each one there is a certain number of VNF (Virtual Network Function). VNF will be ignored in this project beacause are not study objective. This network represent a graph connected component and our goal is to minimize the links weight between two different domains. In fragment_thrinks.py file a simulated net is created. This script generate test.dzn file containing all relative net features. In domain_weights_generation.py file a bunch of scenarios is simulated. This process create ten weight configurations for the same net topology. Output:

• fragment_thrinks.py script give in output simulated.net.dzn file
• domain_weights_generation.py script give scenarios_data.dzn file

These file will be the input files in Minizinc model.

• Python (at least 2.7.13 version)
• Minizinc package

To try this project need to install Minizinc package. Arfet you have runned fragment_thrinks.py and domain_weights_generation.py files, you can run from MiniZinc IDE the Stochestic_solver.mzn file or run from temrinal

mzn-g12mip <file.mzn> <file.dzn>

You can also try to run main_bash.sh script. This script run all file and give you the result.

Enrico Valguarnera
Davide Allevi
Cataldo Picciarelli