EVALUATION OF SERVICE-ORIENTED ARCHITECTURE AND MICROSERVICES THROUGH COMPLEXITY AND PERFORMANCE COMPARISON USING SERVICE GRAPH MODELING

Abstract

Modern enterprise applications use modular and scalable structures, fuelling SOA and microservices
talks. Both architectures rely on loosely connected services, but their structural and operational
distinctions must be considered. It addresses the absence of empirical comparisons between
application performance and architectural complexity. Main goals include a measurable paradigm
for systematic architecture review and data-driven comparison to inform business application design.
A SOA and microservices Vehicle Management System (VMS) application is examined using case
studies. The Service Graph (SG) paradigm describes services and their interdependencies using
graphs. This paradigm employs nodes for services and edges for dependencies. These graphs
calculate architectural measurements like total, global, and coupling variables like CS and RCS.
Apache JMeter simulated 500 and 1000 users for both implementations to assess performance.
Business request (BR) response times were compared for both designs under similar and diverse
service procedures. Despite their higher architectural complexity due to more services and
interdependencies, microservices-based applications respond faster than SOA-based systems,
especially under strong user demands. The SOA method was simple but slow with more loads.
Microservices reduced the response time in professional questions with similar service calculations,
confirming their value in scalable cloud -finance. This was true despite the requests of counting of
separate service. Despite their complexity, Microservice promoted performance in dynamic and
scalable settings, according to the study. The service graph model is effective for architectural
evaluation and can be applied to various systems for studies. This research provides a copyable
method for comparative architectural studies and indicates micro-services infection decisions from
unbroken or SOA-based applications.