Performance improvement of the Data Acquisition System to support the observation quality system of the ASTRI Mini-Array

The ASTRI (Astrofisica con Specchi a Tecnologia Replicante Italiana) program was born as a collaborative international effort led by the Italian National Institute for Astrophysics (INAF) to design and realize, within the Cherenkov Telescope Array (CTA) framework, an end-to-end prototype of the Small-Sized Telescope (SST) in a dual-mirror configuration (2M). The pro
totype, named ASTRI-Horn, is operative, being installed at the INAF observing station located on Mt.Etna (Italy). The ASTRI project includes the building of a Mini-Array of nine ASTRI telescopes that will be installed and operated at the Teide Observatory (Spain). The ASTRI software supports the operations of the ASTRI-Horn prototype and, eventually, of the Mini-Array.

The data acquisition system of the ASTRI-Horn telescope acquires, packet by packet, the read-out data from the back-end electronics of the ASTRI camera. The packets are then stored locally in one raw binary file as soon as they arrive. During the acquisition, the packets are grouped by data type (scientific, calibration, engineering) before processing and storing the decoded data in FITS format. A quick-look component, running on the same machine, that allows the operator to display the decoded data during the acquisition, was also implemented.

During software testing and operations performed with the ASTRI-Horn prototype, due to a cftisio writing method we experienced a bottleneck in the raw-to-FITS binary data conversion when the acquisition rate was greater than about 3 Mbps. We thus decided to apply a workaround and to postpone the conversion in FITS format of the ASTRI-Horn raw data after the end of the acquisition run.

The ASTRI Mini-Array software requires monitoring the data quality as soon as data are available, and this evaluation is performed by the online observation quality system (OOQS). For this reason, and because the Mini-Array consists of 9 ASTRI cameras observing in parallel, we are evaluating a solution to directly send the decoded data from the data acquisition system to the OOQS, strictly reducing any bottleneck.

In this paper, we will present a solution to improve the data transfer efficiency from the data acquisition system, based on redis, to the OOQS component of the ASTRI Mini-Array. Results of the comparison between the acquisition software developed for the ASTRI-Horn and this solution based on Redis will be also shown and discussed.


Theme – Data Processing Pipelines and Science-Ready Data