StarChain

StarChain is a decentralized research-data network for astronomy and space science, gathering observations scattered across observatories, universities and institutes into one searchable storage layer. Its DataStore indexes each dataset by collection, wavelength band and time slice, so researchers can run joint queries across collections, while raw FITS images and metadata are checked for integrity on write. Against the large volumes, mixed formats and high cost of cross-institution reuse, the network applies a common naming and version record so an observation cited by several papers can be traced back to its source.

Storage contribution, retrieval requests and compute jobs are priced separately: nodes are settled by the space and bandwidth they actually occupy, requesters pay by query complexity, and compute tasks are bid out to nodes willing to take them. The companion StarChain AI uses neural networks to classify star catalogues and light curves and to forecast their time series, writing results back to the dataset side for reuse. In August 2024 the project announced the ingestion of roughly 500 TB of astronomical observations and validated cross-dataset queries with researchers at Leiden University; the catalogue lists radio and infrared holdings under the same metadata rules.