Key Takeaways from the 2018 Gdevops Global Agile Operations Summit in Beijing

Introduction

On September 21, 2018 the Gdevops Global Agile Operations Summit opened in Beijing, offering a high‑quality technical program that combined operations and database expertise. The event was recorded and the key sessions are summarized below.

Main Hall – AIOps and DevOps

Song Hui (Senior Manager, Newju Network Operations) emphasized that 2018 marked the “year of AIOps implementation.” He outlined three recommendations for traditional enterprises: unified data management with visualization, AI‑enabled automation, and standardized operational scenarios, forming a roadmap for intelligent operations.

Pei Dan (Associate Professor, Tsinghua University) presented an academic perspective on “unmanned operations,” proposing an AIOps‑based “operations brain” that decomposes complex tasks into four modules, builds an AIOps architecture, decision algorithms, and a knowledge graph to achieve truly autonomous operations.

Ma Zhijie (General Manager, JFrog China) described a three‑step approach to adopt DevOps: select appropriate tools, collect DevOps metadata to evaluate development efficiency, and implement internal containerization. He also stressed integrating vulnerability scanning as a quality gate and using dependency management to clarify impact scopes.

Shen Jian (CTO, KuGou Taxi) focused on managing underlying technology teams, identifying five pain points and offering solutions such as aligning work demand with product development, shifting support from ticket‑based to business‑oriented, establishing clear data and process metrics, and fostering technical sharing sessions.

Database Track

Li Dan (Chief DBA, Luojia Thinking) gave a practical guide to mastering MongoDB, covering read‑write consistency, deployment recommendations, common pitfalls in the latest authentication mode, and query‑optimization principles to help DBAs and developers use MongoDB effectively.

Deng Shuan (Core Engineer, PingCAP) demonstrated how TiDB, a cloud‑native database, can be automatically deployed and operated on Kubernetes, illustrating the evolution of TiDB into a cloud‑native service and the steps to manage stateful workloads on K8s.

Wang Lei (Senior Data Architect, China Everbright Bank) discussed graph‑database applications in banking, such as credit management and anti‑money‑laundering, and provided selection criteria for mainstream graph databases.

Wang Tao (Co‑Founder & CTO, SequoiaDB) explained the architecture of the NewSQL database SequoiaDB, its MySQL‑compatible layer, distributed features, performance testing, and real‑world banking use cases, including strategies for active‑active deployment.

Zhang Liang (DBA Lead, Xiaomi Operations) introduced the open‑source SQLOptimizer tool, detailing its heuristic rewriting algorithm, data‑sampling methods, and index‑optimization strategies for complex SQL queries.

Wang Bin (Project Lead, Cetus at NetEase) described the Cetus MySQL middleware architecture, covering performance tuning, SQL feature extensions, code and compression optimizations, and hardware support.

Operations Track

Zhang Xiaohu (Senior Director, China Telecom Sweet Orange Finance) shared a multi‑site active‑active solution, highlighting traffic scheduling, message distribution, and data‑layer redesign to overcome latency and consistency challenges across distant data centers.

Liu Hua (Technical Manager, HSBC Fund Services) identified common obstacles when applying DevOps in banking projects—complex organizational structures, unclear requirements, and vendor dependencies—and offered practical remedies such as traditional planning tools, agile communication practices, and requirement concretization.

Nie Xin (Operations Director, Tencent) presented Tencent’s “Weaving Cloud” monitoring system, explaining multi‑dimensional data processing, DPL metrics, and four AI‑driven stages: NLP for text, predictive analytics, information convergence, and root‑cause analysis.

Liang Xiaoping (Senior Architect, JD.com) detailed the “Archimedes” cluster management and scheduling system, focusing on container ecosystem, memory over‑commit algorithms, SLA metrics, and leveraging big‑data compute resources for large‑scale promotions.

Xu Xinlong (Senior Operations Engineer, Ctrip) outlined Ctrip’s AIOps practice across anomaly detection, intelligent alert diagnosis, hybrid deployment, and auto‑scaling, emphasizing the need for massive time‑series data to enable fault prediction.

Zhang Juan (Senior Operations Engineer, Alibaba) revealed Alibaba’s NoOps elastic capacity hosting, describing capacity planning, elastic scaling, and risk control measures that dramatically improve availability beyond traditional “nine‑s” targets.

Conclusion

The summit showcased a broad spectrum of operational innovations—from AI‑enhanced monitoring and autonomous operations to cloud‑native database deployments and advanced DevOps practices—providing attendees with actionable insights for modernizing enterprise IT infrastructure.