Mastering the One‑Second Rule: Boost Mobile User Experience in 2024

1. Introduction

Liu Xin, Deputy General Manager of Tencent Wireless Operations, introduces his role supporting tool‑type products such as App Store, mobile browser, phone manager, PC manager, and Tencent Maps.

He shares a story about the launch of the mobile game "Love to Eliminate" on WeChat (iOS) and Mobile QQ (Android), where the Android version suffered low download conversion due to CDN and network issues.

Analysis revealed that download success depends on terminal type, network conditions, and CDN performance. A quick temporary fix changed CDN scheduling from domain‑based proximity to IP‑based proximity, improving metrics.

The story illustrates the need for end‑to‑end user‑experience management in mobile internet services.

2. Mobile Internet Characteristics and the “One‑Second Rule”

Mobile internet traffic passes through multiple network layers: handset → base station → controller → carrier core → internet → cloud services. Unlike wired networks, a mobile device must first acquire a wireless radio channel before any data can be transferred.

Key characteristics:

High latency and low bandwidth (2G: 200‑600 ms, 3G: ~600 ms, 4G: 10‑20 ms, 5G: <1 ms).

Limited radio resources managed by complex signaling; resources are shared among users and reclaimed when idle.

Variable DNS resolution, protocol overhead, and connection establishment times.

The “One‑Second Rule” defines the maximum acceptable time for critical steps (DNS lookup, connection establishment, server response, page rendering) to keep perceived latency under one second on different network generations.

2G: DNS and connection < 1 s.

3G: First‑byte time < 1 s.

Wi‑Fi/4G: First‑screen rendering < 1 s.

3. How to Deliver a Good User Experience

The optimization approach is divided into three layers:

Access‑scheduling optimization

Protocol optimization

Business‑logic optimization

3.1 Access‑scheduling Optimization

Reduce DNS impact by using multiple domain names, dynamic IP lists, and client‑side measurements to select the nearest CDN node. Maintain an accurate IP‑to‑region database (≈98 % accuracy abroad, ≥97 % domestically) to route users to the fastest entry point.

Implement client‑side speed tests to choose the optimal access point for each network switch (4G, Wi‑Fi, etc.), avoiding repeated DNS lookups.

3.2 Protocol Optimization

Key principles include disabling TCP fast‑recovery, setting initial RTO to ≥3 s in weak networks, enlarging the initial congestion window for sub‑1 MB pages, reusing connections, limiting concurrent connections, and compressing headers and payloads.

3.3 Business‑logic Optimization

Adjust long‑connection heartbeat intervals, minimize unnecessary network calls, and batch messages with intervals > 1 s. Reduce UI‑related power consumption by optimizing screen rendering and network usage.

4. Optimizing User‑Experience Metrics

Define quality standards (SLR, SLO) for page‑open time, download success rate, installation success, and activation rate. Collect end‑to‑end metrics via SDKs (“Magic Mirror”, “Lighthouse” projects) and enforce release gates based on these standards.

Iterative releases have raised download completion from 96 % to 99 % and reduced average startup time, keeping 80 % of requests under one second.