How Cloud‑Native Fuels Operations Digital Transformation – Insights from China Mobile

Speaker Background

Wang Xiaozheng is General Manager of China Mobile (Zhejiang) IT and Data Management Department, a senior engineer, chief expert of China Mobile Group, Oracle‑certified database specialist, and honorary mentor of the TGO Kunpeng Association.

1. Cloud‑Native Operations Transformation Challenges

Before 2008, China Mobile focused on voice, SMS, and data traffic. After 2019, the company pursued connectivity, compute, and capability, aiming for digital transformation and high‑quality development. Two essential elements emerged: data and technology.

Without new technical architecture, cost, efficiency, and quality cannot be balanced or amplified. Technology underpins data; without cloud‑native upgrades—such as cloud‑native databases, containerization, and micro‑service architectures—digitalization remains a hollow concept.

The speaker illustrated a military‑style equation (1000‑2×500) to convey that cloud‑native’s fragmentation and distribution dramatically increase operational load, making traditional ops unsustainable.

In cloud‑native environments, the number of network elements grows exponentially, leading to inevitable failures. Current cloud‑native stability lags behind legacy IOE systems, especially as domestic‑made solutions accelerate, rendering pure resource‑and‑person‑based ops infeasible and driving the need for digital ops transformation.

2. Thoughts on Digital Operations Transformation

What is the essence of ops digital transformation? It is “reciprocal strike”: replace manual, human‑centric operations with compute‑driven, machine‑to‑machine interactions, turning data and computing power into the core of ops.

Is AIOps a panacea for traditional industries? While large‑model AI (e.g., ChatGPT) relies on massive public data, ops scenarios involve private data, making large‑model adoption premature; small‑model breakthroughs may exist but lack disruptive impact.

Should telecom enterprises adopt DevOps? The speaker argues against it: without matching talent, organization, and culture, DevOps can produce 1 + 1 < 2 outcomes. Instead, ops should become “development‑like” while development should not become “ops‑like”.

3. Zhejiang Mobile’s Ops Digital Transformation Practice

The practice revolves around six "可" capabilities:

可预测 (Predictable) : Shift from reactive fault handling to proactive fault elimination.

可灰度 (Gray‑scale Release) : Achieve full cloud‑native, AZ‑aware applications and centralized control; results include 80% of requests deployable via gray release, 95% of bugs fixable online, 50% reduction in SRE turnover, and 100% of developers no longer required on‑site.

可观测 (Observable) : Emphasize business‑level observability; sharing data across teams is key, while over‑emphasizing root‑cause analysis may not always restore services.

可协同 (Collaborative) : Even with future autonomous operations (L4/L5), human collaboration remains essential for handling complex incidents.

可逃生 (Escape) : Leverage multi‑region active‑active disaster recovery to embed expert knowledge into runbooks, enabling automatic L3‑to‑L4 escalation and faster business recovery.

可守底 (Baseline Protection) : Implement BASE‑theory‑based “ark” mechanisms as a secondary channel to automatically take over critical business workloads during severe failures.

4. Outlook

Future Vision 1 – Time Dimension: Continuously strengthen ops scenario construction, expanding digital immunity and autonomous driving from isolated pilots to core domains, moving from 1‑5‑10 milestones to a unified 1‑1‑1 model.

Future Vision 2 – Space Dimension: Advance metaverse‑enabled collaborative digitalization, exploring distributed, cross‑domain cooperative ops to reduce multi‑domain coordination costs and achieve cost‑effective, high‑efficiency joint operations.