Reported impact
Business outcome
G-Town completed a full architecture upgrade in two months and gave its AI digital platform a more elastic and controllable runtime. Resource utilization rose, idle spend fell, releases sped up from hours to minutes, and the team improved stability for bursty business events. The serverless stack also made it easier to keep iterating on digital services tied to customer-lifecycle optimization and supply-chain decision support.
- Cost Reduction
- Productivity Improvement
- Scalability
- Speed to Market
Solution
G-Town rebuilt the runtime for its AI digital platform on Alibaba Cloud around Serverless Application Engine (SAE) and the cloud-native API gateway so services could scale and release faster without heavy infrastructure management. Business services are deployed to SAE, where load-driven scaling handles short peaks and can shrink idle workloads to one instance, while CPU Burst and Dragonwell JDK reduce Java startup overhead. The CI/CD pipeline was refactored around kubectl-sae, YAML-defined service specs, and one-click deployments so packaging, upload, rollback, and gray rollout could be automated instead of managed node by node. In front of those services, the MSE gateway provides routing, security controls, rate limiting, and gray-release policy management, while monitoring and log services improve diagnosis. The result is a serverless application platform that underpins G-Town's machine-learning-driven digital services while reducing operational drag.
Data flow
Business requests for G-Town's AI digital platform enter through the MSE gateway, which applies routing and traffic policies before sending them to SAE-hosted application services. SAE scales those services according to load, while monitoring signals and logs flow into Alibaba Cloud observability services so releases, incidents, and high-traffic events can be diagnosed and tuned more quickly.
