Roadmap¶

Phase 1 — Core Reliability (done)¶

• Базовый pub/sub
• Publisher retry task (100ms check, 10s timeout)
• Publisher NACK handling (immediate resend)
• Server sliding window per subscriber
• Server retry task (500ms check)
• poll_fn connection loop (TCP deadlock prevention)
• Zero-copy serialization
• NotifyChannel
• Peak: ~2.45 GB/s end-to-end

Phase 2 — Security & Production¶

• TLS/mTLS (rustls или native-tls)
• Certificate-based authentication
• ACL для очередей (кто может писать/читать)
• Configuration system (TOML/YAML, env vars, CLI overrides)
• Structured logging (tracing)
• Prometheus metrics

Phase 3 — Advanced Features¶

• Message persistence для Durable/Invulnerable на стороне сервера
• Dead Letter Queue (сообщения после N retries)
• Priority queues (High/Normal/Low)
• Web dashboard для мониторинга
• Request-Response pattern для RPC между shards

Phase 4 — Scale & Performance¶

• Server clustering (gossip protocol или shared state)
• Connection pooling на клиенте
• UDP/realtime mode (без acks, для VOIP/games)
• Compression (LZ4/zstd для больших сообщений)
• Topic patterns с wildcards (shard.*, shard.1.events)

Открытые TODO в коде¶