The creation of an international shielding benchmark database was presented in 1988 at the International Reactor Shielding Conference (ICRS7) in Bournemouth, UK. M. Salvatores was among the authors of the proposal and had promoted and contributed to the project since the first initiatives, showed continued interest and encouraged the development of the database. He was Chairman of the Committee of Reactor Physics (NEACRP) for 2 years (1984-1985) and Chair of the Shielding benchmark group (1982-1988). In particular, he chaired two annual meetings in 1984 and 1985, called to initiate the collaborative programme on the analysis of shielding benchmarks for the validation of the JEF data files where the need to organize shielding benchmark was recognized and the presentation at ICRS7 defined the overall project.. SINBAD officially started in the early 1990’s as a collaboration between the OECD/NEADB and RSICC with the goal to preserve the information on the performed radiation shielding benchmark experiments and make these available in a standardised form to the international community. One key point concerned the sensitivity and uncertainty analyses required to define their quality and figures of merit. The database comprises now 102 shielding benchmarks, divided into three categories, covering both low and inter-mediate energy particles applications: fission reactor shielding (48 benchmarks), fusion blanket neutronics (31), and accelerator shielding (23) benchmarks. The database is intended for different users, including nuclear data evaluators, computer code developers, experiment designers and university students. SINBAD is available from RSICC and from the NEA Data Bank. The database was extensively used within the scope of numerous national and international projects, such as PWR Pressure vessel surveillance, fusion programme (ITER reactor studies), different OECD Working Parties on Evaluation Cooperation (WPEC) Subgroups, nuclear data validation, IAEA nuclear data projects, etc. The history of the database and few examples of its use are illustrated, for cross-sections, response functions and covariance matrix validation.