Architecture Overview

This document delves into the internal design of SocketMeister, outlining how it achieves high throughput, resilience, and flexibility.

1. Core Components

  • SocketClient

    • Manages outbound TCP connections to one or more endpoints.
    • Implements automatic reconnection with exponential backoff.
    • Supports round‑robin and failover across multiple servers.

  • SocketServer

    • Listens on a specified IP and port (IPv4/IPv6).
    • Accepts concurrent client connections using a thread‑safe queue.
    • Raises MessageReceived and RequestReceived events for application logic.

  • SocketSession (internal)

    • Encapsulates a single TCP socket and associated state.
    • Handles send/receive loops on dedicated threads.

2. Threading & Concurrency

  • Dedicated I/O Threads

    • Each active session runs its own send/receive loops to avoid blocking.
    • Background worker threads manage connection monitoring and reconnection.

  • Synchronization

    • Internal queues and event dispatch use thread‑safe collections (`ConcurrentQueue`).
    • Minimal locking ensures high concurrency with low contention.

3. Connection Management

  • Automatic Reconnection

    • On disconnect, the client enters a retry loop with configurable backoff intervals.
    • Callbacks allow applications to monitor connection state.

  • Load Balancing & Failover

    • When multiple endpoints are provided, each send selects the next active session.
    • If a session fails, traffic shifts seamlessly to remaining endpoints.

4. Message Framing & Protocol

  • Length-Prefixed Frames

    • Messages are sent with a 4‑byte length header, ensuring correct reassembly.
    • Supports streaming of large payloads without partial message delivery.

  • Request/Response Pattern

    • SendRequest and SendResponse methods handle synchronous round‑trip calls.
    • Timeouts and cancellation tokens ensure callers aren’t blocked indefinitely.

5. Optional Compression

  • Built-in GZip Compression
    • Toggleable per session or per message.
    • Ideal for text-heavy or large binary payloads to reduce bandwidth.

6. Extensibility & Customization

  • Custom Serializers

    • Pass your own encoder/decoder delegates for JSON, Protobuf, or proprietary formats.

  • Pluggable Logging

    • Expose events for raw byte sends/receives to integrate with your logging infrastructure.

7. Performance & Scaling

  • Benchmarks

    • Achieves thousands of messages per second with minimal GC pressure.
    • Pooling of buffer arrays reduces memory churn.

  • Deployment Scenarios

    • Single instance apps to scaled-out microservices farms behind load balancers.

For detailed APIs and configuration options, see the API reference. For code examples, check out the Getting Started guide and our samples folder.

Edit this page