Step-by-step tutorials that take you from zero to a running workload on a 3-node Ferrosa cluster. Each guide starts from docker compose up and ends with live queries.
| Before you begin: Every tutorial in this collection uses the same 3-node Docker cluster. If you haven’t set one up yet, start with the 3-Node Cluster Setup guide. It takes about 5 minutes. |
Getting Started
- 3-Node Cluster Setup
-
Install Docker and launch a production-like Ferrosa cluster in 5 minutes. The directory also includes
docker-compose.mixed-clients.ymlfor deployments that need host/public CQL addresses and separate in-network/internal addresses from the same cluster. - Cluster Scaling: Development to Production New!
-
Watch Ferrosa scale from a single development node to a 3-node preview cluster while writes and schema changes continue during each transition.
Data-Intensive Workloads
- IoT Sensor Data
-
Ingest millions of sensor readings with time-series tables.
- Real-Time Analytics
-
Aggregate financial market data for dashboards.
- RRD Time-Series Aggregation New!
-
Roll up high-frequency sensor readings into time-window summaries (min/max/avg) automatically — RRD-style consolidation with streaming built-ins and optional WebAssembly aggregates.
- Advanced Aggregation
-
Group-by aggregates with built-in functions and WebAssembly user-defined aggregates — including writing scalar UDFs as inline AssemblyScript that the server compiles at definition time.
- E-Commerce
-
Product catalog, shopping cart, and order processing.
- Messaging & Chat
-
Store and retrieve conversation history at scale.
Security & Compliance
- Fraud Detection
-
Score transactions in real time with pattern matching.
- Cybersecurity Monitoring
-
SIEM-style event logs with fast threat lookup.
- Healthcare Records
-
Patient timelines with temporal versioning.
Industry-Specific
- Content Personalization
-
Recommendation engine with user preference tracking.
- Telecommunications
-
Call detail records and real-time billing.
- Gaming Leaderboards
-
Player stats, rankings, and match history.
Vector Search & AI
- Vector Indexes: HNSW and HVQ New!
-
Semantic similarity search over embeddings, comparing the full-precision HNSW index with the quantized HVQ index — plus a BTree secondary index — all from plain CQL.
Multi-Model Search
The lens that makes Ferrosa not just Cassandra: one dataset queried five ways — keyword (LIKE), meaning (vector ANN), fuzzy name (phonetic), trending (RRD), and influence (graph) — and the results fused into one ranked list. No five-system stack, no ETL.
- Multi-Model Scholarly Search New!
-
A paper search engine over one corpus: keyword (LIKE) + vector ANN over real
nomic-embed-text-v2-moeembeddings, phonetic author lookup, RRD citation-trend rollups, and a Cypher citation graph — composed with reciprocal-rank fusion. - Multi-Model Threat Hunting New!
-
A threat-intel workbench over one indicator corpus: keyword over IOC notes, vector ANN over real behaviour embeddings, phonetic actor-name lookup, RRD activity-spike rollups, and a Cypher attack-infrastructure graph — fused into a single triage order.
Ferrosa-Exclusive: Graph + CQL
- Graph Joins & Traversals
-
Relationship queries that traditional Cassandra can’t do — using Ferrosa’s built-in graph engine.
- Research Knowledge Graph
-
Academic papers, citations, authors, institutions, and funding — a full knowledge graph with 42 Cypher queries.
Compatibility & Reference
- CQL Comprehensive Compatibility
-
Systematic test of every CQL feature: all data types, collections, UDTs, tuples, counters, static columns, LWT, batch, TTL, TIMESTAMP, secondary indexes, and more.