SDN ネットワーク技術の普及に伴い，ネットワーク技術者の需要が高まり，技術者育成の演習が多くの教育機関で行われている．実機を利用する演習では，人数分の演習機器を用意する必要などの問題が生じていたため，信州大学ではこの問題を解決するために，ハードウェア仮想化を用いて，効率的な演習を行っている．しかし，SDN 演習では，学習者ごとに演習環境をハードウェア仮想化を用いる手法で提供していたため，演習環境の提供数に限りがあるという問題が生じていた．そのため本研究では，コンテナ仮想化によって SDN 演習環境を構築することで，学習者が Docker・Open vSwitch・trema といった技術を利用できる環境の軽量化を行った．その結果，ハードウェア仮想化と比較し，2 倍程度のメモリ効率を示した．

With the spread of SDN network technology, the demand for network engineers has been increasing, and many educational institutions have been conducting exercises to train engineers. In order to solve this problem, Shinshu University has developed a hardware virtualization system. In order to solve this problem, Shinshu University is using hardware virtualization for efficient training. However, in the SDN training, since the training environment was provided for each learner using the hardware virtualization method, the number of training environments was limited. However, in SDN training, the number of training environments provided to each learner is limited because of the hardware virtualization method. In this study, we propose to build SDN training environments using container virtualization. In this study, we constructed an SDN training environment using container virtualization to lighten the environment where learners can use technologies such as Docker, Open vSwitch, and trema. As a result, the memory efficiency was about twice as high as that of hardware virtualization.