On-chain apps that handle lots of data (games with frequent moves, social feeds, IoT streams) hit a wall: putting everything on the settlement layer is slow and expensive. Data Availability (DA) layers solve this by providing cheap, high-throughput data posting while settlement stays on Cardano or EVM. EffectStream now supports Celestia as a first-class data availability layer.
Part 1 walked through the Celestia integration as an architecture — the batcher, the funnel, the database, the demo UI. This second post is the document-the-subtleties counterpart: when should you reach for a DA layer at all? Both Avail and Celestia are first-class options in EffectStream — same primitive shape, different chains under the hood. Choosing between them, or choosing to skip the DA layer entirely, is the first design decision a team needs to make. After that, the second decision is how to index the data you posted, because a blob on a DA layer is useless to your application until something reads it back into PostgreSQL.