Skip to content

Latest commit

 

History

History
70 lines (51 loc) · 5.61 KB

ADR-20-Architecture-Style.md

File metadata and controls

70 lines (51 loc) · 5.61 KB
Raw

ADR 20: Event-Driven Architecture with Microservices for ClearView System

Date:

2024-09-30

Status:

Accepted

Context:

The ClearView recruitment system aims to provide a robust, scalable, and high-performing platform that integrates with multiple HRMS systems while leveraging AI LLMs for resume recommendations, candidate scoring, and job matching. The following architectural characteristics are critical for the system:

  1. Scalability – The system must dynamically handle varying loads, including fluctuating numbers of candidates, job postings, and integrations.
  2. Performance – Low latency and fast processing are required for real-time AI-driven recommendations, data fetching from HRMS systems, and user interactions.
  3. Interoperability – The system needs to seamlessly integrate with various HRMS platforms, supporting smooth data exchange and maintaining flexibility for future integrations.

Decision

We have decided to adopt an architectural style that is primarily event-driven, with microservices as supporting components for ClearView's recruitment system. This hybrid architecture will provide a flexible and modular approach to meet the system's core requirements.

Justification

1. Event-Driven Architecture (EDA)

An event-driven architecture is chosen as the primary architectural style for the following reasons:

  • Scalability

    • Dynamic Workload Handling: The system can scale by independently scaling event consumers during peak usage, such as when there is a surge in resume submissions or AI processing tasks.
    • Decoupled Components: Services operate independently and communicate via events, allowing individual scaling without impacting other parts of the system.

  • Performance

    • Real-Time Processing: EDA supports near real-time processing of events, critical for low-latency resume parsing, candidate scoring, and job matching.
    • Optimized Resource Utilization: Resources are used more efficiently since tasks are processed only when events are triggered, reducing unnecessary load.
    • Minimized Latency: Asynchronous event processing helps avoid latency associated with synchronous API calls, especially for data synchronization across HRMS platforms.

  • Interoperability

    • Seamless Integration with HRMS Systems: Events facilitate communication between ClearView and various HRMS platforms, standardizing data exchange and processing across disparate systems.
    • Extensibility for Future Integrations: New HRMS integrations can be easily added by subscribing to relevant events without disrupting existing workflows.

2. Microservices Architecture

Microservices are used to complement the event-driven approach for the following reasons:

  • Scalability

    • Independent Service Scaling: Each microservice can be scaled independently based on specific workload requirements (e.g., scaling AI services for more intensive processing).
    • Service-Oriented Scalability: Services can adopt different scaling strategies, optimizing resources based on their usage patterns.

  • Performance

    • Optimized Execution: Services can be fine-tuned and deployed independently, allowing for targeted performance improvements.
    • Asynchronous Communication: Services interact asynchronously through events, reducing blocking operations and improving system responsiveness.

  • Interoperability

    • Standardized APIs for Integration: Microservices expose well-defined APIs for external interactions, facilitating consistent data exchange with various HRMS platforms.
    • Support for Multiple Integration Protocols: The architecture can accommodate different communication protocols, enhancing flexibility and compatibility.

Consequences

Positive Impacts

  • Modular and Resilient Design: The system can evolve, scale, and recover independently due to the decoupling of services.
  • Enhanced Workflow Orchestration: Event-driven patterns allow orchestration of complex workflows (e.g., integrating AI scoring, fetching HRMS data, and updating candidate records) with flexibility.
  • Efficient Error Handling and Recovery: The architecture supports robust error management through event retry mechanisms and fallback strategies.

Negative Impacts

  • Increased Architectural Complexity: The combination of EDA and microservices adds complexity to the system, requiring careful design and monitoring.
  • Eventual Consistency: Due to asynchronous processing, there may be scenarios where data consistency is not immediate, potentially requiring mechanisms to handle stale data.
  • Management Overhead: Managing distributed services, event streams, and message queues introduces additional operational overhead.

Alternatives Considered

  1. Monolithic Architecture: Rejected due to scalability limitations and the lack of modularity needed for integrating multiple HRMS platforms and AI components.
  2. Pure Microservices Architecture: Rejected because it lacks the real-time event-driven processing capabilities needed for low-latency AI recommendations and flexible workflows.
  3. SOA (Service-Oriented Architecture): Rejected as it does not provide the same level of decoupling and real-time processing efficiency as EDA combined with microservices.

Decision Outcome

The event-driven architecture with microservices approach has been approved for the ClearView recruitment system. This architecture aligns well with the requirements for scalability, performance, and interoperability, ensuring a robust and adaptable system.