From IOE to Cloud: China Mobile Zhejiang’s Journey to Modern Ops

Preface

The author, deputy general manager of China Mobile Zhejiang’s IT department and head of the Cloud Computing Center, shares the company’s cloud‑operation practices.

1. Our IOE Era

In the late 1990s the carrier adopted HP high‑end mini‑computers (IOE) and later built an Oracle‑centric database team in 2003. Since 2011 the IOE stack has been rapidly replaced, impacting both telecom and finance services.

2. Background and Drivers for Leaving IOE

To meet the flexibility, elasticity and openness demanded by 4G services, the carrier needed to transform its architecture despite its strong, traditional foundation.

The business growth and the need for distributed, internet‑scale capabilities also required a shift in cost control and social responsibility.

3. Conservatism vs. Progress

Changing a long‑standing system is difficult; the article uses the historical analogy of English longbowmen versus Chinese archers to illustrate the challenge of abandoning legacy technology.

4. The Cloud‑Transformation Journey Begins

Research on cloud began in 2009, and by 2013 the core databases were migrated to X86 servers, completing the IOE removal.

4.1 Our Cloud Path

From 2011 the core CRM system ran on X86; by 2015 all core systems were on X86 servers. Docker was introduced in 2014, and by June 2016 the province‑wide CRM front‑end was fully containerized on DC/OS, with ongoing migration of all core systems to containers.

4.2 Challenges Ahead

First , cloud migration must not compromise stability and availability; standards are even higher.

Second , the technology stack change heavily impacts the team.

Third , cost savings and cost stability must be maintained during transformation.

Fourth , the team’s identity—whether it remains a pure operations team—is questioned.

The team has shifted from a traditional operations group to a DevOps‑oriented platform development team, building a new cloud platform architecture.

4.3 Embracing Cloud Transformation

The company is moving from heavyweight ITSM to lightweight ITSM, redefining the operations team as a platform builder and architecture governance group rather than a passive IOE maintainer.

5. New Achievements in the New Model

5.1 New Operations Organization

The pure operations team has been infused with development capabilities, creating full‑stack engineers that bridge system maintenance.

5.2 More Transparent Issue Handling

Traditional operations suffered from poor automation, visibility and efficiency, leading tenants to distrust the platform. By exposing status proactively and providing visual tools, tenant perception improved.

Automation now includes automatic database anomaly handling and disaster‑recovery switching via a mobile app.

6. Entering a New Battlefield

The team has transitioned from system maintenance to cloud platform planning and construction, adopting a “research‑prototype‑test‑promote” workflow to avoid siloed technology decisions.

Core systems such as mobile business hall and CRM front‑end now run entirely in the cloud, and the core databases have been fully migrated to X86 servers, achieving the IOE‑free goal.

7. Practicing to Ensure Stability

Disaster‑recovery drills once reached 300 per year; with the new architecture the frequency has decreased, but the success rate remains above 99.9%, and switching can be performed via a mobile app within minutes.

8. Facing Tomorrow’s Uncertainties

Key concerns include defining service‑level objectives for private‑cloud tenants, demonstrating IT and cloud platform value to business units, and managing rapid technology turnover without destabilizing the team.

9. Epilogue

The author draws an analogy to historical transformation, expressing confidence that operations teams can reinvent themselves and thrive beyond the “scapegoat” role.