The challenge of the data in the SKA Regional Centres network

Andrea Possenti · 2025

Read The challenge of the data in the SKA Regional Centres network by Andrea Possenti (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03103.txt.

Introduction

Andrea Possenti describes in 2025 the data challenge facing the world's largest radio telescope network. At full operation the observatory will hand researchers over seven hundred petabytes a year. Analysis will happen in a web of regional centres that also curate and archive both observatory products and user-made advanced products.

The centres follow principles that keep data findable, accessible, interoperable, and reusable, and they help researchers estimate computing needs when proposing projects. Possenti reports the network's status and outlook, with special attention to the Italian node led by the national astrophysics institute. The piece frames data handling, not telescopes, as the bottleneck ahead.

Catalog record

CREATORAndrea Possenti
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE15 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-03103

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The challenge of the data in the SKA Regional Centres network Andrea Possenti Abstract At regime, SKAO is expected to provide the researchers with an annual amount of more than 700 hundred PB of data. The advanced analysis of all those data will take place within a network (SRCnet) of so-called SKA Regional Centres, which, under the ‘Findable, Accessible, Interoperable, and Reusable’ (FAIR) principles, will also take the responsibility for curating and archiving both the Observatory Data Products and user-generated Advanced Data Products, as well as for helping the researchers in the estimate of the needed computational effort at the proposal stage. This contribution describes the status and the perspectives of this international network, with particular emphasis on the INAF-led Italian node which is part of that. 1 The SKA Observatory (SKAO) basics SKAO is the largest radio telescope facility ever conceived. It will be built in two very under-populated sites: (1) one (dubbed SKA-LOW) will be located in the Murchison area in Western Australia and will comprise 131072 log-periodic dipole antennas, grouped in 512 stations of 256 antennas each, spread across a maximum baseline of 74 kilometers and having a frequency range between 50 and 350 MHz; (2) another one (dubbed SKA-MID) will be located in the Karoo area in South Africa and will include 133 dish-like antennas of 15 m in diameter, combined with other 64 antennas of 13.5 m in diameter, which are part of the already operating Meerkat radio telescope [ 1 ] . The initial frequency range will be split across three observational bands, the first two running from 0.35 to 1.1 GHz and from 0.95 to 1.76 GHz, while the third covering the 4.6-15.4 GHz range. In the final configuration, additional bands will be implemented and the maximum baseline will reach about 150 kilometers. The (3) third component of SKAO, the Headquarters, are located close to the historical, but still active, Jodrell Bank Observatory, nearby Manchester, in the UK. After about 3 decades during which the telescope was designed and its funding and governmental structure assessed, SKA is now in construction at the two sites, adopting a staged approach, which will lead to a first fully functional system by around the end of 2027 and 2028. At the website [ 2 ] full details can be found about the telescope configuration, the transformative science which it will deliver, the development phases, the Countries contributing to the project (16 Countries are currently involved in that as fully fledged members or at a national-coordination level or are represented as observers) and the organizational structure. 1.1 The key role of computing in the SKA project Among the experimental sciences, radio astronomy is historically one of the most demanding in terms of data acquisition, storage and computing power necessary for the full exploitation of the collected data. In fact, radio astronomy observations require sampling and digitizing the analog signal that is received and amplified by the so-called antenna front-end. This sampling takes place both in the time domain and in the frequency domain. In many kinds of traditional observations, the band of frequencies observed is split into thousands or, sometimes, in tens of thousands of channels, and each of these channels is sampled at time intervals ranging from once per second in imaging observations until, in some time-domain observations, to more than 10000 samples per second. More recently, technology advancements allowed to also routinely collect data from an antenna up to the Nyquist frequency associated with the selected width of the observing band (the so-called baseband recording). While this is in principle the optimal choice (no information is lost in the data acquisition), it imposes very highly demanding requirements in terms of produced data: for instance, observing over a band with a width of 800 MHz (a rather common choice in modern radio astronomy) implies to sample the data at least at 1.6 GHz, translating into a data rate of 1.6 GB per second if one wants to digitize the analog signal at 8 bits. With these cadences, the flow of data coming out from each antenna of an array of antennas (or from each station of the array) can become enormous. In the specific case of the SKAO, the design of the antennas and receivers systems is expected to produce an internal flow of data of the order of 2 PB per second, in turn pushing the requirements for the signal processing in the range of the 10^{18} multiply–accumulate operations (i.e. the exa-MAC range) and the associated needed high performance computing enters the ballpark of the hundreds of Petaflop/s. The ingestion and analysis of these data will occur in quasi-realtime at the two telescopes’ sites, mainly in two large data processing machines, called CPF (Central Processing Facility) and SDP (Science Data Processor). As a result, at regime, it is expected that about 700 PB of the so-called Observatory Data Products (ODP) will be generated by the SKAO. As matter of fact, the output data rates from SKAO are extraordinary for a science experiment: of order 50-70 times higher output than other big astrophysical future facilities, like the Vera Rubin Telescope (aka LSST, operating in the optical band) and the CTAO (operating in the TeV regime). The data rate from SKAO will be comparable to that of the High Luminosity LHC experiment at CERN. As such, it is clear that in the case of SKAO the signal processing capabilities play a fundamental role in making this an unprecedented radio astronomical instrument and are at the heart of the transformative science which this new international facility promises. Indeed, in some fields of the radio astronomical research the signal processing capability of the currently available technology are the limiting factor for achieving even more exciting scientific outcomes with SKAO. If this is something to live with at the beginning, it also opens a great future for this facility, since the growing of the digital technology is expected to occur faster than any additional development of the antennas and front-ends. Thus, besides establishing as a unique instrument since the beginning of its post-commissioning life, the SKAO is due to also improve its capability in a relatively short timescale. 2 The case of the SKA Regional Center Network (SRCnet) As said above, \sim 700 PB of Observatory Data Products will be produced by SKAO in a year. However, in most cases they will not be ready yet for a full analysis and for the subsequent papers’ writing by the scientists. Along the path already followed by other astrophysical projects, the Observatory Data Products will flow over the internet and will be distributed across a number of so-called SKA Regional Centers (SRCs), where they will be archived, cured and additionally manipulated by the scientists in order to produce the so-called Advanced Data Products (ADP). In view of this scheme, on July 2016 the Board of SKA (at the time provisionally leading the project) deliberated that The SKA Observatory will coordinate a network of SKA Regional Centres (SRC) that will provide the data access, data analysis, data archive and user support interfaces with the user community . To make the story short, during the subsequent years, a lot of in-kind efforts were done by

Read the complete text