After the first light, Subaru telescope produced about 86,000 frames or 400 giga bytes data during its test observation by the end of February 2000. STARS (Subaru Telescope ARchive System) contains all data and is serving them to the observers. STARS also provides several convenient tool and information such as QLI (Quick Look Image) by the aid of QP (QLI Producer) and QLIS (QLI Server), HDI (HeaDer Information file), and machine readable (on-line) memorandum for observed data, for making users know the rough quality of the data at a glance. QLI file is a FITS file with FITS BINTABLE extension. By the combination of QP and QLIS (our code name is 'GYOJI'), users have data with various size (20 to 200 times smaller than original one) on their needs, and also many extracted information such as mean, maximum and minimum count values, profiles of extracted spectra in multi-slit spectroscopy or echelle spectroscopy data and so on in the original data browser (QLISFITS) written as JAVA2 applet. This functions will also be used for public data archive system in the future. For the convenience of the data analysis, STARS also handles and manages the 'dataset,' which is essential for preparing the necessary data including object and calibration frames used in data analysis by DASH (Distributed Analysis System Hierarchy: platform for data analysis of Subaru Telescope data). This 'dataset' is made at the summit system (SOSS: Subaru Observing Software System) which knows everything about the procedure of the observation performed, and is interpreted by DASH system. In this paper, we will describe the functions which STARS has and how STARS, DASH and SOSS are linked each other for leading the effective scientific and engineering returns.