Modern scientific applications often require provisioning of computing and networking infra-structures tailored to particular application needs which might not be clearly defined. In addition, allo-cated resources of an actual physical infrastructure might be underutilized by applications (e.g., a pro-cess does not utilize provided CPU or network connection fully). Traditional virtualization technolo-gies can solve the problem partially, however, lead to noticeable overheads. In this paper we consider creation of virtual topologies to test distributed software behaviour without the need to construct real network topologies. We evaluate an approach to create dedicated computing environments configured for particular applications based on light-weight virtualization also known as containers. We investi-gate available capabilities to model and create dynamic container-based virtual infrastructures sharing a common set of physical resources, and evaluate their performance on a set of test applications with different requirements.