This describes an engineering demo showing Zero’s next-generation IVM engine squeezed into a single Durable Object (DO) and used to query a 100GiB database stored in Cloudflare R2 while keeping live queries active. The experiment, posted by Matt (@tantaman) on Oct 16, 2026, reports serving a 100.6 GB dataset from one DO with a live demo at demo.r2db.dev. The stated goal was to find the smallest, most constrained runtime able to run the IVM, and the result is a one-DO deployment that reads from R2 and maintains live query state.
The practical takeaway is that a full incremental-view-maintenance engine can be extremely compact and still handle large edge-hosted datasets with live updates, demonstrating a path to low-footprint, real-time database services at the edge. Packing the engine into a single durable runtime reduces architectural complexity and suggests cost and latency benefits for real-time apps that rely on edge storage and continuous queries. The post generated visible community interest, signaling enthusiasm for tiny, powerful edge runtimes serving substantial persistent data.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.