The study aims to systematically review literature on the rare diseases information system to identify architecture of this system from a data perspective. The search for relevant English language articles, based on keywords in title, abstract, Mesh and Emtree terms, was done in Pubmed and Embase (from 1980 to June 2017), Scopus, Science Direct and Cochran (from 1980 to July 2017). Articles were selected if they addressed data architecture of information systems with a focus on rare disease, and if at least one of their objectives dealt with design, implementation, and development of rare diseases information systems. Thirty-five studies met the inclusion criteria. The findings were categorized into six groups. This first group addressed organizations acting as data generators, data users, and data governors. The second group was related to data sources and databases. Datasets and data elements formed the third group of findings, including common datasets, specific datasets, and complementary datasets. The fourth group of findings was in relation to data standards. Data sharing and interactions among relevant bodies included the fifth group of the findings. The last group of findings was pertinent to procedures and criteria used for checking the quality of data, as cross review checking was a main procedure assessing the accuracy, consistency, and completeness of data. Design and development of an integrated information system for rare diseases considering data architecture principles in practice could help eliminating issues with management of rare diseases through facilitating sharing information and experiences.