ultra-low-latency 开源项目推荐（30 个）

ant-media /Ant-Media-Server

Ant Media Server — Ultra-low latency streaming engine with WebRTC (~0.5s), SRT, RTMP, HLS, CMAF, adaptive bitrate, transcoding & scaling

OvenMediaLabs /OvenMediaEngine

OvenMediaEngine (OME) is a Sub-Second Latency Live Streaming Server with Large-Scale and High-Definition. #WebRTC #LLHLS

libriscv /libriscv

The fastest RVA23 RISC-V sandbox

morganstanley /Xpedite

A non-sampling profiler purpose built to measure and optimize performance of C++ low latency/real time systems

krish567366 /submicro-execution-engine

Sub-microsecond bare-metal execution engine with deterministic replay, lock-free order path, and hardware-timestamped latency measurement.

hwangsaeul /gaeguli

[gæguli] is SRT streamer designed for edge devices that require strong security and ultra-low latency.

sashamakarenko /fixpp

FIX library generated with C++ PreProcessor to be used in latency sensitive context.

coralblocks /CoralPool

CoralPool is a high-performance, lightweight and garbage-free Java object pool implementation. It efficiently reuses mutable objects, minimizing the creation of short-lived objects that would otherwise burden the garbage collector.

gonewx /nextalk

nextalk - High-performance offline voice input for Linux 实时语音识别系统

mrkingsleyobi /quantumedge

⚡ Ultra-low-latency HFT platform: FPGA (14ns FIX), GPU TensorRT (<1ms ML), lock-free, DPDK. 350ns latency, 500K orders/sec. Production-ready. 99.99% uptime.

coralblocks /CoralDS

Garbage-free, single-threaded data-structures (maps, sets and lists) optimized for speed.

coralblocks /CoralAffinity

Comprehensive and straightforward Linux thread affinity in Java so that you can pin your critical threads to isolated CPU cores for maximum performance.

giangndm /decentralized-p2p-streaming

Minimize server usage by leveraging a decentralized peer-to-peer network for ultra-low-latency live streaming among users.

varnish /libvmod-riscv

Sandboxing and multi-tenancy for Varnish

userFRM /crossbar

Ultra-low-latency zero-copy pub/sub over shared memory. O(1) transfer at any payload size. URI-addressed, multi-publisher. Apache-2.0.

ucloud /ubox-publisher

RTMP stream publisher with hardware accel

adanil-code /OptimisticCache

An allocation-free, ultra-low-latency C++20 concurrent cache for multi-platform User-Mode and the Windows Kernel (WDM/KMDF). Utilizes SeqLocks and NUMA sharding to deliver wait-free, nanosecond reads.

ryntric /channels-java

It is a low-latency Java concurrency library designed around ring buffers, sequencers, and customizable wait strategies.

Praiven /P4-SFU

Bachelor Thesis investigating if an SFU implemented in P4 can drastically reduce latency in multi-party conferencing systems.

ryntric /affinity-dispatcher

Affinity Dispatcher is a high-performance, low-latency message routing framework in Java. It routes messages to workers based on key affinity and supports multiple key types.