The DNI (Direct Normal Irradiance) is the resource utilized by solar concentrators such as in CSP (Concentrated Solar Power) and CPV (Concentrated Photovoltaic) applications for electricity production or in solar bakery applications in off-grid remote communities. The main goal of the SoCooMa (Solar Concentration for Cooking in Mafate) project is to provide a detailed assessment of Surface Solar Radiation (SSR) availability for concentrating solar bakeries in Reunion Island with a focus on the isolated area of Mafate. The accurate instrument that can be used for DNI measurements is the pyrheliometer. However, this instrument is expensive and requires high maintenance efforts, as it must be mounted on an expensive automatic two-axis sun tracker to point precisely at the sun all day. Since the measurements of GHI (Global Horizontal Irradiance) and DHI (Diffuse Horizontal Irradiance) from the cost-effective SPN1 pyranometers at the Indian Ocean Solar network (IOS-net, https://galilee.univ- reunion.fr/) stations have good accessibility and availability over Reunion Island, the present paper aims at clarifying to what extent these measurements can be used for deriving the DNI time series for the SoCooMa project. It also explores the possibilities provided by satellite-based DNI databases for places where no pyranometer measurements exist. Additionally, the paper presents a study of the estimation procedure for DNI from 1-minute experimental data of GHI and DHI from IOS-net in Saint-Denis (Réunion Island University, Moufia campus site), as well as the evaluation of satellite- based DNI at the same site. As the DNI in this study is indirectly derived, an error analysis is then performed for both DNI values (SPN1 and satellite) using the reference measurements of the Kipp&Zonen CHP1 pyrheliometer from Reunion BSRN station at Saint-Denis. Finally, correction procedures are applied to the DNI raw data of the SPN1 and satellite in order to reduce their deviation from the reference DNI determined by the CHP1 (pyrheliometer).