Igor Suhorukov: A knowledge base without a second database: Oxigraph’s SPARQL in PostgreSQL 19

Oxigraph's integration of SPARQL directly into PostgreSQL 19 eliminates the need for a separate triple‑store database, allowing developers to query knowledge graphs using standard SPARQL without managing a second data store. For website owners, this means that any on‑site knowledge base—product catalogs, skill ontologies, or FAQ datasets—can now be served from the same PostgreSQL instance that powers the website. The reduced architecture simplifies scaling, cuts latency for AI‑driven search features, and lowers the attack surface that bots might exploit when probing multiple endpoints. However, the unified endpoint also concentrates traffic, so AI crawlers and malicious bots targeting SPARQL queries will now hit the primary database directly, potentially affecting site performance and data security. **Actionable tips:** 1. **Update your llms.txt** to include a `Disallow: /sparql` rule for known AI crawlers (e.g., `User‑Agent: ChatGPTBot`). This prevents automated bots from hammering the SPARQL endpoint while still allowing human users. 2. **Implement query‑rate limiting** at the PostgreSQL level (using `pgbouncer` or built‑in connection limits) specifically for the `/sparql` path, throttling requests from IP ranges identified as bots. 3. **Monitor SPARQL traffic** with your existing bot‑tracking dashboard; flag spikes in query complexity or volume and automatically trigger alerts or temporary blocks via your firewall rules.