How China Telecom’s Dual Systems Earned SRE Level‑3 Certification and Elevated SOMM Operations

With rapid digital‑technology updates, information systems have become critical, and their stability faces new challenges. The Chinese government’s Key Information Infrastructure Security Protection Regulation (effective 1 Sept 2021) mandates operators to ensure safe, stable operation, prompting many enterprises to adopt Site Reliability Engineering (SRE) practices.

Assessment Result

On 10 July 2026, the China Academy of Information and Communications Technology (CAICT) announced the 38th batch of dual‑certificate IT‑DevOps assessments. China Telecom’s “Product‑Sales Management System” and “Enterprise‑Marketing Control System” both passed the CAICT SOMM SRE Level‑3 evaluation, indicating a marked upgrade in the company’s operational maturity.

Systems Overview

The CPCP system links front‑end business with back‑office services, managing ~80 000 sales items, 1 500 products, and serving >5 000 active users across the group. The enterprise‑marketing control system handles 14 key business nodes, aggregates multi‑source data, and supports >30 000 users with >1 000 concurrent daily sessions.

Interview Highlights

Motivation for Assessment – Align with industry standards, obtain a clear benchmark, and use the evaluation to drive systematic stability improvements and share experience with the industry.

Benefits Observed – Team perception shifted from “tools and methods” to a “culture and organizational capability.” The assessment reinforced proactive prevention, pre‑emptive drills, and data‑driven decision‑making. Annual availability now exceeds 99.9% with zero recorded incidents.

Key Capability Enhancements

Full‑stack observability covering infrastructure, middleware, application, and business services.

Routine chaos‑engineering experiments to validate resilience.

Capacity planning with predictive, dynamic scaling.

Automation of operations from manual checks to platform‑based, visualized workflows.

Operational Improvements – Monitoring alerts now catch >90% of issues before customers notice, reaching SRE maturity Level 4. Automated incident response reduced MTTR from hours to minutes.

Challenges Faced – Unifying SRE baselines across two systems with different ages and tech stacks. CPCP (launched 2014) required observability retro‑fit, while the newer marketing system needed robust chaos‑engineering and capacity planning.

Solution Strategy – Apply a unified SRE framework with system‑specific metric thresholds, address legacy observability gaps, and build a common operations platform to reduce duplicated effort.

Five Core Practices

Metric‑driven: Implement SLI/SLO/error‑budget tracking.

Automation‑first: Shift manual tasks to automated pipelines.

Continuous chaos drills: Regular fault injection in test and prod.

Post‑incident retrospectives: Root‑cause analysis within 48 hours.

Cultural leadership: Foster a shared reliability responsibility across dev, ops, and test teams.

Future Plans – Pursue Level‑5 maturity, integrate AI‑based anomaly detection and automated root‑cause analysis, expand SRE coverage to more core systems, deepen AIOps‑SRE integration using eBPF, and strengthen talent development for hybrid SRE engineers.

Industry Outlook – SRE will evolve from observability to predictability via AI, adopt self‑healing mechanisms, and tightly bind reliability metrics to business value, turning reliability into a competitive asset.

Key Visuals