Identifying Alternatives and Deciding Factors for a Data Mesh Architecture

Abstract

The data mesh was introduced in 2019 as a new type of data architecture. It promises a more democratic and scalable way of data production and consumption, while also solving data engineering problems of siloed and hyper-specialized data engineering knowledge, a growing number of dependencies within data pipelines, and the rigidness of centralized monoliths. This paper used expert interviews to identify the most significant current alternatives to the data mesh and abstract factors, with which companies can evaluate whether a data mesh can further their move to a data-driven, democratized future. The results show that the motivation, company culture, company structure, IT history and IT structure should be evaluated before implementing a data mesh. This paper is based on a bachelor thesis.