HeliosProxy Feature Roadmap
HeliosProxy Feature Roadmap
Vision
Transform HeliosProxy from a basic connection router into the most intelligent, AI-ready database proxy for modern applications. By combining traditional proxy capabilities with innovations for AI/ML workloads, agentic systems, and distributed caching, HeliosProxy will become an indispensable layer between applications and HeliosDB-Lite clusters.
Strategic Pillars
1. Performance at Scale
Connection pooling, query caching, and distributed caching to handle millions of requests with sub-millisecond latency.
2. Intelligence & Adaptability
Self-tuning routing, workload-aware optimization, and ML-driven query analysis.
3. AI/Agent Native
First-class support for AI agents, RAG pipelines, vector operations, and agentic workflows.
4. Developer Experience
REST APIs, GraphQL gateway, query hints, and comprehensive observability.
5. Enterprise Ready
Multi-tenancy, authentication proxy, rate limiting, and compliance features.
Feature Categories
Phase 1: Foundation (Core Proxy Enhancements)
| # | Feature | Priority | Complexity | Status |
|---|---|---|---|---|
| 1 | Connection Pooling Modes | Critical | Medium | Planned |
| 2 | Query Caching | Critical | High | Planned |
| 3 | Query Routing Hints | High | Low | Planned |
| 4 | Replica Lag-Aware Routing | High | Medium | Planned |
Phase 2: Resilience & Control
| # | Feature | Priority | Complexity | Status |
|---|---|---|---|---|
| 5 | Rate Limiting & Query Throttling | High | Medium | Planned |
| 6 | Circuit Breaker Pattern | High | Medium | Planned |
| 7 | Query Analytics & Slow Query Log | Medium | Medium | Planned |
Phase 3: Enterprise Features
| # | Feature | Priority | Complexity | Status |
|---|---|---|---|---|
| 8 | Multi-Tenancy Support | High | High | Planned |
| 9 | Authentication Proxy | Medium | High | Planned |
| 10 | Query Rewriting | Medium | Medium | Planned |
Phase 4: Differentiation & Innovation
| # | Feature | Priority | Complexity | Status |
|---|---|---|---|---|
| 11 | WASM Plugin System | High | High | Planned |
| 12 | GraphQL-to-SQL Gateway | Medium | High | Planned |
| 13 | Schema-Aware Routing | Medium | High | Planned |
Phase 5: Distributed Intelligence (Flagship Feature)
| # | Feature | Priority | Complexity | Status |
|---|---|---|---|---|
| 14 | Helios-DistribCache | Critical | Very High | Planned |
Timeline Overview
Q1 2026: Phase 1 - Foundation├── Connection Pooling Modes (transaction/statement)├── Basic Query Caching├── Query Routing Hints└── Replica Lag-Aware Routing
Q2 2026: Phase 2 - Resilience├── Rate Limiting & Throttling├── Circuit Breaker└── Query Analytics
Q3 2026: Phase 3 - Enterprise├── Multi-Tenancy├── Authentication Proxy└── Query Rewriting
Q4 2026: Phase 4 - Innovation├── WASM Plugin System├── GraphQL Gateway└── Schema-Aware Routing
Q1-Q2 2027: Phase 5 - Helios-DistribCache├── Core distributed cache├── Workload specialization├── Heatmap & analytics└── AI/Agent optimizationsAI/Agent Integration Philosophy
Every feature is designed with modern AI workloads in mind:
| Workload Type | Proxy Optimizations |
|---|---|
| RAG Pipelines | Vector similarity caching, embedding prefetch, chunk locality |
| AI Agents | Session affinity, conversation context caching, tool call routing |
| Agentic Workflows | Transaction batching, idempotency keys, retry intelligence |
| LLM Applications | Response caching, semantic deduplication, streaming support |
| Vector Search | HNSW index-aware routing, ANN result caching |
HeliosDB-Lite Synergies
Each proxy feature exposes and amplifies HeliosDB-Lite capabilities:
| HeliosDB-Lite Feature | Proxy Enhancement |
|---|---|
| WAL Streaming | Lag-aware routing uses real-time lag metrics |
| Sync Modes | Route critical writes to sync standbys |
| Vector Search | Cache ANN results, route to vector-optimized nodes |
| TWR (Transparent Write Routing) | Extend to proxy level for any PostgreSQL client |
| Multi-Model Storage | Route JSON/vector/relational queries to specialized nodes |
| Columnar Storage | Route OLAP queries to columnar-optimized replicas |
Success Metrics
Performance
- Connection efficiency: 100x improvement with transaction pooling
- Query latency: <1ms for cached queries
- Cache hit ratio: >90% for hot data
Reliability
- Failover time: <100ms automatic failover
- Circuit breaker: Prevent cascade failures
- Zero downtime: Rolling updates without connection drops
Observability
- Query insights: Full query analytics without pg_stat_statements
- Cache heatmaps: Visual cache utilization
- Latency histograms: P50/P95/P99 per query pattern
Getting Started
Each feature document includes:
- Overview: Problem statement and solution
- Architecture: Technical design
- API Specification: Configuration and endpoints
- AI/Agent Innovations: Modern workload optimizations
- HeliosDB-Lite Integration: Synergies with core database
- Implementation Notes: Development guidance
Browse individual features in the features/ directory.
Contributing
Feature proposals welcome! Each proposal should include:
- Problem statement
- Proposed solution
- API design
- Performance implications
- AI/Agent use cases