EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Compound Agentic Intelligence for Legacy Modernization: A Reasoning-Orchestrated Framework for Agentic AI and Generative AI

Arun Meesala

International Journal of Scientific Research in Science, Engineering and Technology · 30.11.2024 · 11 (6) · 623-632

DOI: 10.32628/ijsrset2512557 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Scientific Research in Science, Engineering and Technology, online erschienen am 30.11.2024. Verfasst von Arun Meesala. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Enterprise legacy modernization—migrating monolithic applications to microservices architectures—remains labor-intensive, requiring manual dependency analysis, service boundary identification, and continuous integration reconfiguration. Generative AI coding assistants accelerate individual development tasks, yet operate as stateless suggestion engines disconnected from deployment pipelines, container orchestration, and cross-environment validation, leaving modernization dependent on engineer-driven coordination across disparate tools. This paper proposes the Compound Agentic Modernization Framework (CAMF), a reasoning-orchestrated architecture fusing large language model (LLM) reasoning with deterministic build, test, and deployment systems to autonomously drive legacy-to-microservices migration. CAMF introduces four original components: a Service Boundary Reasoning Agent (SBRA) decomposing monolithic codebases into candidate microservice boundaries via dependency-graph analysis, a Migration Plan Synthesizer (MPS) generating phased refactoring sequences validated against deterministic build constraints, a Pipeline Orchestration Agent (POA) autonomously configuring CI/CD stages across build and deployment systems, and a Cross-Environment Validation Layer (CEVL) verifying functional parity across development, QA, UAT, and production environments. The framework was evaluated against twelve enterprise .NET monolith-to-microservices migration projects comprising 4.2 million lines of code. CAMF reduced manual migration effort by 64.3%, decreased post-migration defect density by 71.8%, and shortened average pipeline configuration time from 18.4 to 3.1 engineer-hours per service. These results indicate that compound agentic architectures, combining probabilistic reasoning with deterministic verification, offer a viable path toward trustworthy autonomous modernization at enterprise scale.

Zitieren

APA

Meesala, A. (2024). Compound Agentic Intelligence for Legacy Modernization: A Reasoning-Orchestrated Framework for Agentic AI and Generative AI. International Journal of Scientific Research in Science, Engineering and Technology, 11(6), 623-632. https://doi.org/10.32628/ijsrset2512557

Vancouver

Meesala A. Compound Agentic Intelligence for Legacy Modernization: A Reasoning-Orchestrated Framework for Agentic AI and Generative AI. International Journal of Scientific Research in Science, Engineering and Technology. 2024;11(6):623-632. doi:10.32628/ijsrset2512557

BibTeX
@article{Meesala2024,
  author = {Meesala, Arun},
  title = {Compound Agentic Intelligence for Legacy Modernization: A Reasoning-Orchestrated Framework for Agentic AI and Generative AI},
  journal = {International Journal of Scientific Research in Science, Engineering and Technology},
  year = {2024},
  volume = {11},
  number = {6},
  pages = {623--632},
  doi = {10.32628/ijsrset2512557}
}
Kostenlos starten →