Containerization for creating reusable model code

Abstract

Will you be able to run your computational models in the future? Even with well-documented code, this can be difficult due to changes in the software frameworks and operating systems that your code was built on. In this paper we discuss the use of containers to preserve code and their software dependencies to reproduce simulation results in the future. Containers are standalone lightweight packages of the original model software and their dependencies that can be run independent of the platform. As such they are suitable for reuse and sharing results. However, the use of containers is rare in the field of modeling social-environmental systems. We provide an introduction to the basic principles of containerization, argue why it would be beneficial if this tool became common practice in the field, describe a conceptual walkthrough to the process of containerizing a model, and reflect on near future directions of containerization workflows.

OPACA: Toward an Open, Language- and Platform-Independent API for Containerized Agents

“…As described in [23], modularity is an important design principle for software design. Concepts such as Docker images also have shown that modularity often comes along with reusability [24]. 4) Self-Descriptive: It should be self-describing in the sense that the components using the API provide information about their functionality and how to use them.…”