How Gome Used Cloud Computing & Automation to Revolutionize IT Ops

Today’s topic is Gome Group’s cloud computing and operation automation practice. Although Gome is traditionally known as a large home‑appliance retailer, the company has embraced cloud technologies and automated operations to improve efficiency and reduce costs.

Gome Overview

New Operation Thinking

Cloud Computing Practice

Operation Automation Practice

1. Gome Overview

Founded in 1987, Gome operates about 1,500 stores and employs roughly 300,000 staff, making it the largest domestic home‑appliance chain. Beyond retail, Gome is involved in six major sectors: real‑estate development, internet finance, smartphones, intelligent manufacturing, internet services, and offline retail.

The technology side supports a wide range of business lines, requiring robust infrastructure. Since around 2000, Gome has unified order storage and built a hybrid environment that includes both traditional IOE and cloud/virtualization components.

2. New Operation Thinking

Gome’s operation strategy focuses on two core goals: reducing operation costs and increasing operational efficiency. To achieve this, the company leverages cloud computing and automation.

In June 2016, Gome established a dedicated subsidiary, Gome Cloud, to provide IAAS services for all internal business units, eliminating duplicated infrastructure investments.

By centralizing resources such as cabinets, servers, and network equipment, Gome creates a shared resource pool that can be allocated on demand, avoiding idle assets and lowering expenses.

Automation is also unified: each subsidiary’s automation requirements are consolidated into a single platform, enabling standardized, large‑scale deployments.

3. Cloud Computing Practice

Gome Cloud’s primary product is IAAS, with plans to layer PAAS and SAAS on top. Because building a private data center is time‑consuming and costly, Gome also integrates third‑party cloud services to meet rapid delivery needs.

The IAAS architecture is divided into three layers:

User Layer : SRE personnel from each subsidiary access the portal to request cloud resources via APIs.

Control Layer : Handles resource scheduling, billing, and backend services.

Service Layer : Includes CMDB, GMSTACK (similar to OpenStack), RDS, SLB, VPC, CDN, VPN, monitoring, and a MIX‑Cloud component that aggregates third‑party cloud APIs.

The control component, called

Command

, is written in Go and orchestrates resource provisioning, supporting synchronous, asynchronous, and concurrent requests via HTTP, HTTPS, and RPC (Thrift).

Monitoring started with Zabbix for thousands of hosts, then transitioned to a customized Open‑Falcon solution that provides fine‑grained metrics (CPU, memory, disk) directly from the hypervisor, reducing noise and supporting alarm aggregation.

4. Operation Automation Practice

Gome’s automation ecosystem revolves around a CI/CD‑style lifecycle that standardizes processes, integrates development, testing, and configuration management, and provides a unified platform for all subsidiaries.

Key automation products include:

CMDB that treats network resources (e.g., IP addresses) as assets and maps users, devices, and services.

Asset‑driven rack and cabinet visualization for efficient hardware provisioning.

A custom PXE/PXD‑based installation system (ROMOS) that automatically assigns IPs, RAID configurations, and performs rapid mass deployments.

Load‑balancing based on LVS and Nginx, supporting HTTP‑level gray‑release by dynamically adjusting server pools.

Integration with Docker for containerized workloads, complementing traditional VM‑based services.

Through these initiatives, Gome has transformed its IT from a cost center into a profit‑generating service platform, achieving lower costs, faster delivery, and a more agile operation model.