tidwall/evio

Features

• Non-Blocking Event Loops - Provides a non-blocking event loop architecture to process network traffic with minimal system overhead.
• Event Loop Load Balancers - Distributes incoming network traffic across multiple internal event loops using round-robin or least-connections strategies.
• Round-Robin Load Balancers - Implements a system for distributing network connections across multiple CPU cores using round-robin scheduling.
• Connection Lifecycle Events - Implements a callback system to manage connection states and data flow during network event transitions.
• High-Performance Networking - Designed for building scalable, low-latency network applications in Go that handle massive concurrent traffic.
• Protocol-Agnostic Interfaces - Provides a protocol-agnostic interface for handling TCP, UDP, and Unix domain sockets via asynchronous callbacks.
• Multi-Threaded Packet Processing - Distributes network traffic across multiple CPU cores using event loops to maximize throughput.
• Go Event-Driven Networking Frameworks - A high-performance Go library providing a non-blocking networking framework for TCP and UDP servers.
• Socket Event Handling - Implements a callback-driven system for processing asynchronous read, write, and error events on network sockets.
• Non-Blocking Socket I/O - Provides a framework for non-blocking socket I/O to handle concurrent connections with minimal latency.
• Multi-Core Scaling - Distributes network processing across multiple event loops to maximize throughput on multi-core systems.
• Custom Protocol Implementations - Provides the low-level primitives necessary for implementing specialized network protocols with fine-grained socket control.
• Socket Buffer Management - Manages network connection lifecycles and data buffers to reduce memory allocation and processing delays.
• Multi-Protocol Server Implementations - Enables the creation of servers that handle both TCP and UDP traffic on a single port.
• Network Socket Bindings - Enables listeners to bind to several network interfaces or socket types simultaneously using a single event loop.
• Port Sharing - Supports socket reuse allowing multiple sockets on a single host to bind to the same port.
• Socket Port Reuse - Supports socket reuse allowing multiple sockets on the same host to bind to a single port.
• Shared Buffer Allocations - Reduces memory overhead by reusing a single input packet buffer across multiple concurrent network connections.
• Packet Processing Frameworks - Utilizes a high-performance event loop mechanism for efficient network packet processing.
• TCP and Unix Socket Listeners - Accepts incoming traffic through both standard TCP ports and local Unix domain sockets.
• Connection State Management - Tracks and manages the active state and lifecycle of network connections via event outcomes.
• Stream Buffer Allocators - Utilizes a specialized memory management strategy that shares a single input packet buffer across multiple connections.
• Multi-Protocol Listeners - Allows a single port to listen for and handle both TCP and UDP network traffic simultaneously.

Star history