Year: 2023
Type: dissertation
Abstract: Legacy systems are still in widespread use, especially in corporate settings and factories. They often use proprietary communication protocols, making it harder to integrate them with new software and third-party applications. Implementing REST APIs for such systems can improve their interoperability and ease of use. When implementing new APIs for legacy systems security is of particular concern. This thesis investigates the security considerations of building a REST API for a legacy system. It does this in the form of a case-study of the development of a secure REST API backend for a Manufacturing Execution System. The thesis investigates how the REST API can be implemented in a secure manner, and which security aspects have to be considered. It also investigates special security considerations for legacy systems. The legacy system in question relies heavily on logic running in the database, which led to a focus on an API that directly interfaces with the database. This choice in turn put a focus on proper validation and authorization procedures. For interoperability, some security relevant systems, such as authentication, were made compatible with the legacy system. For implementing the API a model-driven development approach is used, where the meta-model is based on the OpenAPI Specification (OAS). The meta-model is highly domain-specific in order to incorporate aspects such as authorization and validation. It is based around the concept of database tables and rows being modeled as REST resources. In addition to this, traditional non-model-driven solutions are also used for some security related functionality. The resulting implementation was found to be secure, but there were cases where more secure options than those implemented were available. Particularly, compatibility with the legacy systems required some compromises and special considerations. The use of model-driven development was found to improve the security of the API backend.
Source: Aaltodoc (Aalto University)
Author: Lönnbäck, Otto
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7112519555
