Multi-tenancy: Interview Felipe Montoya

Date: February 19, 2025
Attendees:

• Felipe Montoya (eduNEXT)

• Andrés González (Aulasneo)

• Dave Ormsbee (Axim)

• Guillermo Viedma (Aulasneo) (Facilitator)

Discussion Points:
The meeting was intended to address the following questions:

• Does eduNEXT use Organizations as the base entity for its multi-tenancy implementation? Could you elaborate on that?

• Do you know if other Contributors follow the same structure for implementing multi-tenancy?

• How would you expect the statement "I can create roles in Organization A" to function?

• Should a user from Organization B be able to see the roles created by Organization A?

• Do you think any of these changes could impact your implementation? I am particularly interested in how you handle sensitive data like PII.

• Does eduNEXT use an independent database per Organization to manage multi-tenancy?

Many of these topics were addressed tangentially.

Key Takeaways:

• Felipe Montoya was the primary respondent, while Dave Ormsbee and Andrés González were listeners. Guillermo Viedma facilitated the discussion.

• Felipe explained that eduNEXT was once the champion of multi-tenancy within the community but moved away from this implementation after the introduction of MFE (Micro-Frontend Architecture).

• eduNEXT had invested significant time and development efforts into multi-tenancy as it helped reduce costs for clients. However, after adopting Tutor, implementation costs dropped significantly, making multi-tenancy less necessary. As a result, eduNEXT rarely accepts new clients using this model.

• Felipe believes that the way the platform is currently built makes multi-tenancy unviable. The cost-saving benefits are minimal compared to the high maintenance costs.

• He emphasized that customers often require independent databases, meaning that the challenge is not only with MFE implementation but also with the inability to separate client data in the database, making multi-tenancy implementations less attractive.

• Given the current state of the platform, building features with multi-tenancy in mind is inefficient since very few community members use this implementation. Even eduNEXT, once the primary adopter, is moving away from this model.

• Any changes made should allow platform users to extend functionality as needed, but the cost of implementation should be covered by the company requiring it, not the project's core development, as multi-tenancy is a niche need.

• eduNEXT did not offer deep customization in its multi-tenancy implementations, as access control remained standardized across tenants to maintain cost efficiency and simplicity for clients with budget constraints.

• Felipe considers using Organizations for multi-tenancy unrealistic but acknowledges its usefulness in larger structures, such as universities needing departmental separation for courses.

• Regarding permissions such as "create role," he believes this will likely depend more on a platform administrator rather than a department-level manager.