Architectural Design Activities for JAS

The critical part for building any software system is its architecture. Architectural design is a design at a higher level of abstraction. A good architecture ensures that software will satisfy its requirement. This paper defines the most important activities of architectural design that used through building any software; also it applies these activities on one type of Electronic Commerce (EC) applications that is Job Agency System(JAS) to show how these activities can work through these types of applications.

💡 Research Summary

The paper begins by reaffirming that software architecture is the most critical factor in determining whether a system can meet its functional and non‑functional requirements. It then outlines a structured set of architectural design activities that can be applied to any software project. The activities are grouped into four major phases: (1) stakeholder and requirement analysis, where both functional use‑cases and quality‑attribute scenarios such as availability, scalability, security and performance are elicited; (2) quality‑attribute modeling, which quantifies the importance of each attribute and creates trade‑off matrices; (3) architectural style and pattern selection, where candidates such as layered, service‑oriented, event‑driven and micro‑service architectures are evaluated against the modeled attributes; and (4) component decomposition and interface definition, which produces a domain‑driven component model, explicit service contracts, and a verification plan using methods like ATAM, scenario‑based evaluation and load testing.

To demonstrate the practicality of this framework, the authors apply it to a Job Agency System (JAS), a specific type of e‑commerce application that handles job postings, candidate resumes, matching algorithms, payment processing, and user authentication. In the requirement‑analysis stage, the key actors (employers, job seekers, administrators) and core workflows (post job → match candidates → schedule interview → finalize hiring) are identified. The quality‑attribute model prioritizes 99.9 % uptime, sub‑200 ms response time for matching, strong data privacy, and elastic scalability under traffic spikes.

During the style‑selection phase, a hybrid architecture is chosen: a classic three‑tier presentation‑business‑data stack for most functions, combined with a micro‑service for the matching engine and an event‑driven messaging layer (Kafka) for asynchronous processing. The front‑end is built as a React SPA, protected by an API gateway that enforces OAuth 2.0 and JWT‑based authentication. Business logic runs in Spring‑Boot micro‑services orchestrated by Docker and Kubernetes, enabling automatic scaling and self‑healing. Persistence uses a relational database (MySQL) for transactional data and Elasticsearch for full‑text search, with read‑replicas and sharding to meet scalability goals.

Component design separates domain entities (job posting, candidate profile), matching service, payment service, and notification service, exposing both RESTful and gRPC interfaces to balance interoperability and performance. Validation activities include an ATAM workshop with stakeholders, synthetic load tests simulating 10 000 concurrent users (average latency ≈ 180 ms), and security scans with OWASP ZAP confirming that OAuth 2.0/JWT mitigates common threats.

The results show that the systematic architectural activities enable JAS to satisfy its stringent quality targets while maintaining a clear, modular structure that supports future evolution. The paper concludes that the feedback loops between each activity and the rigorous verification steps are essential for sustaining architectural quality. It also suggests future work on automated quality‑attribute modeling tools and continuous architecture evaluation pipelines to further streamline the design process.