Deployment
QUORUM ships as container images orchestrated with Docker Compose and runs inside the institution's own environment — on-premise or in its private cloud. Deployment profiles select which features run.
Request scanning, honeypots and threat intelligence only.
The rule engine without the AI tiers.
All features except the Kafka bridges.
All features, including the Kafka bridges.
01 // What Runs
- QUORUM application image
- Dataplane sidecar (Go) for kernel and attestation operations
- VOPRF evaluator service for federation
- PostgreSQL and Redis
- Ollama inference for the locally run AI models (GPU recommended)
- Optional Neo4j relationship graph
- OpenTelemetry, Prometheus and Grafana
02 // Failure Behaviour
If the rule engine cannot evaluate a request or the decision cannot be recorded, QUORUM returns no verdict (HTTP 503, ISO 8583 code 91) instead of allowing it.
If the AI tiers are unavailable, real-time decisions continue unaffected; deep-analysis requests return "challenge", or "block" when the institution selects fail-closed.
03 // Not Provided Today
Kubernetes manifests, QUORUM-hosted SaaS, multi-region failover and published capacity tiers are not provided. Latency and throughput figures will be published with their test conditions after measurement in a pilot.