Backend Development 13 min read

Architecture and Practices of 58.com Alliance Advertising Platform

This article presents a comprehensive overview of the 58.com alliance advertising platform, detailing its business model, core modules such as SSP, DSP, media and creative platforms, system architecture, performance optimizations, big‑data processing, and future directions for programmatic creative generation.

DataFunSummit
DataFunSummit
DataFunSummit
Architecture and Practices of 58.com Alliance Advertising Platform

The 58.com alliance advertising platform enables internal advertisers to acquire external traffic through SSP and DSP mechanisms, turning site traffic into revenue by serving ads on partner media.

Business Overview : The platform consists of four main modules—Same‑City SSP, Direct‑投 platform, Alliance DSP, and Creative Platform—each handling media registration, ad placement, external DSP integration, and automated creative generation.

Media Platform : Media partners register companies, media sites, and ad slots, receive placement links, JavaScript or API endpoints, and embed them to monetize inventory. The workflow includes media registration, link deployment, ad request routing to the alliance ADX, ad retrieval from the 58 ad pool, and revenue sharing based on impressions or clicks.

Advertising Platform (MKT API) : Consolidates multiple external ad exchanges (Baidu, Tencent, etc.) to simplify operations. It faces challenges such as reliance on media‑side user profiles, lack of fine‑grained exposure control, and scattered strategy models across DSPs.

DSP Architecture : Handles real‑time bidding (RTB) with a 70 ms external ADX deadline and a 50 ms internal target. To meet 100k+ QPS, the system adopts asynchronous processing with an EventLoop‑Thread model and Future/Promise for callbacks, reducing thread blocking and lock contention.

Performance Optimization – Asynchronous Model : Replaces a blocking thread pool with an event‑driven architecture, improving throughput and latency as shown by benchmark comparisons.

Performance Optimization – Latency & GC : Addresses frequent Full GC pauses caused by a large LRU cache of IP‑to‑city mappings. The solution introduces a fixed‑size three‑dimensional long array (Free GC) and scheduled System.gc() to extend GC intervals from 2 h to 30 h, reducing TP99 latency from 80 ms to 50 ms.

Index Design & OLAP : Uses a Lambda architecture with Flink for real‑time processing and Kylin for offline OLAP. The DSP employs a dual‑index strategy with primary and replica shards, and a two‑level cache for bidding data to support TB‑scale throughput.

Programmatic Creative Platform : Automates creative generation to combat fragmented ad slot specifications and low‑quality assets. The Creative‑Builder engine extracts features, selects templates, applies rule‑based optimization, renders images, uploads to CDN, and indexes them for CTR‑based selection.

Creative Index Routing : Optimizes shard routing by using city + category as a route‑key, achieving 95 % of queries on a single shard, and manages routing tables via Zookeeper.

Future Outlook : Continues to enhance programmatic creative quality, scaling, and routing efficiency while expanding data‑driven personalization.

Performance OptimizationSystem ArchitectureadvertisingBig Dataad techDSPSSPprogrammatic creative
DataFunSummit
Written by

DataFunSummit

Official account of the DataFun community, dedicated to sharing big data and AI industry summit news and speaker talks, with regular downloadable resource packs.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.