PRISM deployment requirements — published transparently with measured benchmarks.
All figures reflect actual production configurations.
Sized for the most common production scenarios.
Baseline: 10 concurrent users running typical OLAP workloads on standard BI datasets.
Concurrent (peak latency may rise) · operating threshold
Concurrent · response headroom · stable ops
Company-wide · multi-department · scale after tuning
PRISM deploys into fully isolated air-gapped networks.
Only three communication paths require firewall configuration — internal container traffic stays within the Docker network.
| Path | From | To | Port / Protocol | Purpose |
|---|---|---|---|---|
| [1] | User PC | PRISM Server | 443 / TCP (HTTPS) | Web Access |
| [2] | PRISM Server | AI Inference Server | LLM API Port / TCP | LLM Inference Call (or external SaaS = 443/TCP) |
| [3-8] | PRISM Server | Customer Source DB | DB Port / TCP | Source Data Query (Oracle 1521, MSSQL 1433, PG 5432, MySQL 3306, DB2 50000, SAP OData 443) |
| [9] | Docker Internal Net · PostgreSQL · Elasticsearch · Redis | No firewall needed | ||
PRISM ships on a single host by default.
When non-stop availability is required, a reference HA architecture separates each tier — preserving service even if a single host fails.
Two app hosts in Active-Active configuration behind an L4 load balancer.
Supports 20 concurrent users with full failover.
PostgreSQL + AGE in Active-Standby with streaming replication.
RPO ≈ 0 · RTO in minutes.