Global Perspective on Multi-Cloud Data Architecture

As China's internet market matures and the population dividend fades, companies are turning to overseas expansion, making efficient and compliant data architectures—especially multi‑cloud and hybrid cloud—critical for success.

In this global arena, cloud services serve as the foundation for digital transformation, enabling businesses to overcome geographic limits. Because regional regulations, market characteristics, and technology adoption vary, a single‑vendor approach is insufficient, prompting the adoption of multi‑cloud strategies that leverage the strengths of providers such as Alibaba Cloud in Southeast Asia, Microsoft Azure in the Middle East, and IBM across the Americas, Europe, and East Asia.

The forum, scheduled for July 6 from 14:00 to 17:50, is produced by Baide (Ji Qianfei), Head of Data Engine at Xiaohongshu, and features experts from Xiaomi, Xide International, Qichacha, Kuaishou, and others sharing technical updates under multi‑cloud architectures.

Talk 1 – Xiaomi’s Lake‑Warehouse Multi‑Cloud Practice : Speaker Xiao Jiebao, a software R&D engineer, will discuss the exploration of lake‑warehouse multi‑cloud architecture, Iceberg‑based multi‑cloud table storage, Gravitino‑driven unified management of non‑tabular Data/AI assets, and future plans. Audience will learn about Xiaomi’s lake‑warehouse architecture, cost‑saving storage solutions, and unified data management.

Talk 2 – Xide International’s Multi‑Cloud Data Platform for Cross‑Border E‑Commerce : Speaker Wang Xiaolei, CTO of Xide International, will describe the need for cloud‑neutral data products, the migration from PostgreSQL to a distributed data warehouse, and the resulting benefits of cloud neutrality, performance gains, development efficiency, and simplified user experience. Attendees will gain insight into cross‑border data platform construction, technology selection, multi‑cloud system building, and scaling strategies.

Talk 3 – Accelerating Cloud‑Based Machine Learning with Alluxio at Xiaohongshu : Speaker Li Yabin, a big‑data technology expert, will present how Alluxio accelerates data reads for ML training, improves indexing and model performance in multi‑cloud settings, and speeds up massive small‑file scenarios. Listeners will learn about data acceleration product selection, intelligent caching for faster training sample reads, and reducing cross‑cloud data copies.

Talk 4 – Qichacha’s Path to Data Cost Reduction and Efficiency : Speaker (unnamed) will outline Qichacha’s data architecture, the formation of a hybrid cloud architecture, and unified multi‑cloud design, highlighting resource efficiency improvements of over 30%, multi‑version architecture, and unified metadata management.

Talk 5 – Kuaishou’s Flink on Kubernetes Migration and Stability Assurance : Speaker Liu Jiangang, a technical expert, will cover background motivations, architectural design across user, platform, and engine layers, stability measures including workflow standards, job management, operational guarantees, and overseas AZ failover, as well as future plans for a dynamic, self‑adaptive, compute‑storage separated real‑time platform. Participants will understand Flink on K8s production practices, stability assurance mechanisms, and the latest industry trends.

The full agenda is available via QR code, and attendees can scan for a discounted 20% group ticket offer.