The Smithsonian dataset is very large. Piri nodes must be able to scale up to store at least 1 PiB of data. This may include setting/recording capacity and taking it into account when routing. Rough design docs here: - https://www.notion.so/storacha/Scaling-Piri-Operations-2ee5305b55248079a18ae7d7d74ba5db - https://www.notion.so/storacha/Scaling-Piri-s-Storage-2ee5305b552480c5a076fe66d2ba95e8