Deep Dive into Vivo Activity System Globalization: Boost Reuse, Cut Maintenance

Background

Vivo’s activity system is a visual, plug‑in based, one‑stop operation platform that enables rapid activity creation through drag‑and‑drop configuration. As the business expands overseas to countries such as Thailand, Singapore, and Indonesia, the need to reuse domestic services in foreign markets becomes essential for globalization.

Internationalization

Multilingual Support

The system originally used simple i18n files that required manual maintenance and release cycles. The upgraded solution integrates a multilingual platform, persisting language data in both the database and a configuration center, providing better performance and user experience. It also offers a common SDK for other services to adopt the same configuration.

Multi‑Timezone Handling

Instead of letting operators manually convert times for each region, the system stores all timestamps in the East‑8 (Beijing) timezone and converts them to the user’s local timezone on display. This reduces both operational and development complexity. For regions with multiple timezones, the default timezone is used with clear activity notes.

Multi‑Tenant Architecture

The platform adopts a two‑dimensional "project‑country" matrix, isolating data, configurations, users, and processes per project and per country. This design yields:

Business Flexibility: New projects or countries can be added by copying a project container.

Fine‑grained Management: Permissions match organizational hierarchy; regional managers see only their countries.

Operational Cost Transparency: Usage, performance, and cost can be drilled down per tenant.

Data is stored in isolated databases with separate credentials, enhancing security, stability, and reliability. Sensitive data is classified and encrypted via a unified proxy.

Global Co‑line (Unified Code Line)

To avoid duplicated effort between domestic and overseas services, Vivo consolidates them into a single code base deployed globally. Two scenarios illustrate the need:

Domestic lottery activity is successful; overseas markets want the same activity.

Same activity type is maintained separately in multiple regions, leading to divergent rules and increased maintenance.

The goal is one code line, unified architecture, and global deployment.

Unified Domain Naming

Requests use a single domain with region code passed via HTTP header (e.g.,

X-I18n-Code : loc=${地区码}; lan=${语言码}; tz=${时区}

). This enables routing to the appropriate datacenter without proliferating domains.

Handling Regional Resource Differences

Three dependency scenarios are considered:

Dependency already globally co‑lined – no extra handling.

Dependency not yet available in a region – either temporary forwarding or reduced functionality, while urging the dependency owner to co‑line.

Dependency varies across regions – use a generic forwarding service that routes calls based on region code.

Post‑Co‑line Service Maintenance

After merging services, every version rollout must be deployed to all regions to avoid code drift, configuration mismatch, and hidden bugs. Although this raises short‑term deployment cost, it reduces long‑term maintenance and improves reuse.

Conclusion

Successful global co‑line requires:

Ensuring existing business stability during migration.

Fully accounting for external dependency variations across regions.

Adopting a global‑first mindset in product, development, and testing to consider all locales in each iteration.

The unified architecture, multilingual and multi‑timezone support, multi‑tenant isolation, and multi‑datacenter deployment together enable high component reuse, lower maintenance overhead, and a scalable global service.