For IoT, along with Cloud, Fog, Dew, Edge computing is important, that provides aggregation and data processing from sensor networks and individual devices related to the physical world. Increasing of productivity of boundary network nodes can significantly reduce data access delays, traffic, and bandwidth of Internet access channels. A lot of network interfaces and protocols IoT make it difficult to develop and debug applications, require the tools for rapid prototyping of network fragments and project-oriented methods for training developers. The approach to the creation of such tools and the allocation of training projects of varying complexity are proposed.