Skip to main content

Overview

Low-Latency Streaming (LLS) is a solution designed for stable, efficient, and scalable live streaming with minimal latency. It provides real-time media processing and an optimized architecture that supports diverse platforms and application-specific low-latency scenarios.

What is Low-Latency Streaming (LLS)?​

LLS addresses the core challenges of live streaming — input delay, processing overhead, and delivery latency — through a purpose-built architecture that optimizes every stage of the streaming pipeline.

Key characteristics of LLS include:

  • Minimal Latency - Edge node connectivity on both ingest and delivery sides reduces end-to-end delay
  • Real-Time Processing - Optimized transcoding and packaging with hardware-accelerated media servers
  • Global Scale - Distributed edge network ensures consistent quality regardless of viewer location
  • Rapid Deployment - Domain and configuration setup can be completed in minutes
  • Flexible Output - Support for multiple delivery protocols and adaptive bitrate configurations

How LLS Works​

LLS adopts an optimized, low-latency architecture that integrates global ingest, high-performance media processing, and scalable content delivery. The system operates through three interconnected layers:

Ingest Layer​

Live streams enter the LLS platform through the nearest edge node in a globally distributed network. By routing streams to the closest ingest point, LLS significantly reduces input latency and improves transmission stability from the source encoder to the processing infrastructure.

Processing Layer​

The ingest edge nodes connect to the core media processing infrastructure through a high-speed backbone network. Media servers equipped with Nvidia GPUs and optimized algorithms perform real-time transcoding and packaging with minimal processing delay. This layer handles:

  • Video transcoding (AVC/H.264, HEVC/H.265)
  • Audio transcoding (AAC, Mono/Stereo)
  • Adaptive bitrate packaging
  • Live recording and screenshot capture

Delivery Layer​

Processed streams are distributed through CDN edge nodes closest to each viewer. The delivery layer routes viewer requests to the optimal edge node, ensuring low playback latency and a consistent quality of experience across all geographic locations.

A GOP (Group of Pictures) cache mechanism at the edge accelerates stream startup time, allowing viewers to begin playback almost instantly upon connecting.

Supplementary Features​

Beyond the core streaming pipeline, LLS provides value-added capabilities throughout the complete streaming lifecycle:

  • Live Recording - Automatically capture and archive live streams for on-demand access
  • Live Screenshots - Generate visual snapshots during live sessions for monitoring and thumbnails
  • Stream Management - Real-time controls to monitor, ban/unban, and reset active streams
  • Analytics & Insights - Comprehensive metrics covering bandwidth, viewers, transcoding, and geographic distribution
  • Anti-Leech Protection - Token-based authentication to prevent unauthorized stream access