Hello Fineract Community, I'd like to seek community feedback on the security model for the Loan Origination System (LOS). Current Architecture
The current LOS prototype is implemented as a standalone service that integrates with Apache Fineract. It currently provides: - A public-facing applicant portal for creating and submitting loan applications. - A staff portal supporting configurable multi-stage approvals (Loan Officer → Credit Committee → Branch Manager). - Authentication using Fineract users and roles. - Loan disbursement through Fineract after the approval workflow completes. Security Considerations One of the areas I am currently evaluating is the security model for a production deployment. The applicant-facing LOS is intended to be publicly accessible, while Apache Fineract Back Office is typically deployed within a trusted internal environment. This raises questions about how responsibilities should be divided between the two systems, particularly for approval workflows that ultimately lead to financial actions. Some possible approaches include: - Keeping both customer submission and staff approvals within LOS, secured with appropriate authentication and authorization. - Keeping customer-facing functionality in LOS while moving approval stages into the Fineract Back Office. - Leveraging the existing Back Office role and permission model for internal approval workflows. The current implementation should be considered a proof of concept, and we intend to clearly document that it is not yet a production-ready security architecture. Questions I would appreciate community feedback on the following: 1. What security model would be recommended for a public-facing Loan Origination System integrated with Apache Fineract? 2. Should multi-stage approvals remain within LOS, or should they be handled by the Fineract Back Office? Thank you. Sujan
