Scalar Supports Tsuneishi Shipbuilding's Core System Modernization with AI-Driven Development; Microservices via Scalar and Kong Foundations

Scalar (Headquarters: Shinjuku-ku, Tokyo; CEO: Wataru Fukatsu), a global leader in data management, and Kong Inc. (Headquarters: Minato-ku, Tokyo; President: Daiki Arisumi), a leader in AI connectivity, announced that Tsuneishi Shipbuilding Co., Ltd. (Headquarters: Fukuyama-shi, Hiroshima; President: Yukio Okumura) adopted 'ScalarDB' and 'Kong Konnect' for its core system modernization project as the first outcome of their collaboration.

In this project, two IT personnel dispatched from Tsuneishi Shipbuilding used AI agents to complete the analysis and refactoring of legacy systems that had remained untouched for over a decade in just two days. The AI-driven development approach succeeded in developing a Minimum Viable Product (MVP) within a short period. The Scalar-Kong collaborative solution robustly supports the flexible microservices environment, ensuring AI code quality.

[Background]

Traditional companies face challenges with 'monolithic' legacy systems that have become bloated and 'black-boxed' over years of expansion, hindering responsiveness to market changes. Tsuneishi Shipbuilding had an urgent need to modernize, yet faced deep-seated issues that had prevented them from even initiating an investigation for over a decade.

While automated system modernization using generative AI and AI agents is gaining traction, processing large-scale systems as a whole often leads to 'context pollution'—where AI generates incorrect code due to excessive information, failing to ensure quality. Success in AI-driven development requires decomposing the system into appropriate business domains (microservices) while simultaneously ensuring 'data consistency' and 'secure API connectivity' across the split systems using a high-level infrastructure.

[Collaborative Solution]

In this project, two IT staff from Tsuneishi Shipbuilding were stationed at Scalar for three months. Under Scalar's technical consultation, they practiced advanced 'microservices migration and AI-driven development' by combining AI agents, ScalarDB, and Kong.

Ultra-fast current state analysis using AI agents (completed in two days)

Analyzing and refactoring complex source code that had been ignored for years was conducted in just two days using AI agents, covering everything from design reconstruction to MVP development.

Microservices architecture to prevent 'context pollution'

To maintain high quality in AI-generated code, the system was migrated to independent 'microservices' for each business domain. By limiting the information AI processes (separating contexts), they achieved high-precision, bug-free development.

Ensuring data consistency and AI agent integration with ScalarDB

ScalarDB's powerful distributed transaction technology was applied to prevent data discrepancies between segmented services. It also functions as a secure data access foundation for AI agents to search and utilize data across multiple microservices.

Instant construction of API management foundation with Kong

'Kong Konnect' was adopted as the API management foundation to securely and rapidly connect segmented services. The collaboration between Kong and Scalar enabled the rapid implementation of the microservices architecture while minimizing infrastructure setup efforts.

[Future Outlook]

With MVP development completed in just three months, the project will now proceed to step-by-step production releases and operation while adjusting with user departments at Tsuneishi Shipbuilding. Furthermore, Scalar and Kong will promote this success story as the 'standard model for AI-driven modernization' and push joint solution expansion and sales collaboration for enterprise companies struggling with core system modernization (DX) and migration, both in Japan and internationally.