You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@fineract.apache.org by mumin iddrisu <mi...@gmail.com> on 2023/11/14 16:42:50 UTC

Self-Service

Hello Team
Has anyone successfully deployed the self-service feature?
If yes please let me know which build you are using.

Regards
Abdul
---mis
ᐧ

Self Service Strategy at Fineract

Posted by James Dailey <jd...@apache.org>.
Community -

I want to highlight a recent discussion.   As Paul writes:  "fineract must
support self-service".
(see emails https://lists.apache.org/thread/qfydfzclwb2wstgqwcgq5hgv7lvm11dr
)

I agree.

However, as previously discussed, the (now removed) implemented
functionality "self service in Fineract" created security issues and was
unsuitable for a production environment.

We have broadly requested a group to step forward and assist.  With no
progress on that, vendors are building their own "plug in".  That is a
natural consequence, although I would prefer to see those contributions at
Fineract.

I have created a GSOC ticket,
https://issues.apache.org/jira/browse/FINERACT-2439, to begin addressing
this deficit in the Fineract software.  The proposal is essentially the
same as it was two years ago - to create a separate component that sits
between the existing Fineract system and the end-users (on apps, on
websites).

This GSOC project is intended to move us forward.

How important is this functionality?  Very.

Where should it sit given our current architecture?  On the fineract
project in a new repo (different component)  and it should use its own
authentication flow.

What should users or deployment teams do today to address this gap?  You
can build your own or use a vendor solution.  If you build your own you can
leverage the data model and API patterns found in the git history.   (just
go back to a commit circa early 2025)   We cannot recommend deploying an
archived release.

What should this look like in the long term?  To be determined by those
working on it.   It is possible that we adjust the data model of Fineract
to accomodate this properly.

I hope that helps.

James

Re: Self-Service

Posted by Ádám Sághy <ad...@gmail.com>.
Hi

A little heads up on this front.

The final nail in the coffin of self-service functionality (inside Fineract) is here:
Story
https://issues.apache.org/jira/browse/FINERACT-2480

PR
https://github.com/apache/fineract/pull/5498

Regards,
Adam

> On May 14, 2025, at 5:18 PM, James Dailey <ja...@gmail.com> wrote:
> 
> Victor - These are important requirements that you raised.  So far, I think some people are on board with these things but I don't yet see a strong effort on this. 
> What would you advise as the sequence and prioritization?   
> 
> How does this relate to the ongoing FSIPs? 
> 
> Do we need to create some tickets to work on?  
> 
> James Dailey
> 
> 
> 
> On Thu, Mar 13, 2025 at 9:19 PM VICTOR MANUEL ROMERO RODRIGUEZ <victor.romero@fintecheando.mx <ma...@fintecheando.mx>> wrote:
>> Hello,
>> 
>> The Apache Fineract APIs must be deprecated because there are more efficient and secure ways to allow end users to access its financial services. 
>> 
>> The Apache Fineract must continue to include financial products on its core modules and it has  to promote the ecosystem around it through interfaces that allow third party systems based on open standards the interchange of information in a secure manner. 
>> 
>> Don't worry there are already plugins (Open sourced which are compatible with the current Apache Fineract version)
>> 
>> A self service APi Architecture and Implementation is not limited to the technologies that I have mentioned before (On Wed, 31 Jan 2024). Grab a cup of coffee and start reading.... 
>> 
>> For a more detailed explanation about how to design a secure Self-Service API REST architecture with access/identity management, OTP, and best-in-class security for the Self Service APIs we must design and implement:
>> 
>> 
>> 1. Authentication & Identity Management
>> OAuth 2.0/OpenID Connect :
>> Use Keycloak, Zitadel, Gluu or Auth0 as an identity provider (IdP) for token-based authentication.
>> Example: Secure endpoints with @PreAuthorize("hasRole('USER')") in Spring Boot.
>> Enforce PKCE (Proof Key for Code Exchange) for public clients (mobile apps).
>> JWT Tokens :
>> Use short-lived access tokens (e.g., 5 minutes or less) and refresh tokens.
>> Validate JWT signatures, issuer (iss), audience (aud), and expiration (exp).
>> Password Policies :
>> Enforce strong passwords (min 12 characters, mix of symbols/numbers).
>> Hash passwords with BCrypt or Argon2 (and never ever store it using plaintext). 
>> 
>> 2. Multi-Factor Authentication (MFA) with OTP
>> TOTP (Time-Based OTP) :
>> Integrate with Aegis, Google, Azure, Authenticator or Twilio Authy.
>> Use Keycloak, Zitadel, Gluu built-in OTP policies for user registration/login.
>> OTP Delivery :
>> Send OTPs via secure channels (e.g., SMS, email, or authenticator apps). 
>> Avoid SMS for high-security scenarios (vulnerable to SIM swaps).
>> Rate Limiting :
>> Limit OTP requests (e.g., 3 attempts per hour) to prevent brute-force attacks using an API Gateway (Kong, Gravitee, ApiSix) .
>> 
>> 3. Authorization & Access Control
>> RBAC (Role-Based Access Control) :
>> Define roles like USER, ADMIN, and SUPPORT with granular permissions.
>> Example: Restrict /self/clients/{id} to the owning user only.
>> Scope-Based Access :
>> Use OAuth2 scopes (e.g., read:profile, write:loans) to limit API actions.
>> Attribute-Based Access Control (ABAC) :
>> Restrict access based on user attributes (e.g., country, client_type).
>> 
>> 4. Data Protection
>> Encryption :
>> In Transit : Enforce HTTPS (TLS 1.3+) using .
>> In Transit: Make sure to use secure CipherSpecs and CipherSuites
>> At Rest : Encrypt sensitive fields (e.g., PII) in the database.
>> Masking Sensitive Data :
>> Log/display masked data (e.g., accountNumber: "CL-****123").
>> Use @JsonView in Spring to control data exposure.
>> 
>> 5. API Security Best Practices
>> Input Validation :
>> Sanitize inputs to prevent SQLi/XSS (e.g., use Hibernate/Apache/Spring Validator annotations).
>> Token Binding :
>> Bind tokens to specific devices/IPs to prevent token theft.
>> 
>> 6. Secure OTP Workflow
>> Step-Up Authentication :
>> Require OTP for sensitive actions (e.g., changing passwords or transferring funds).
>> Secure OTP Storage :
>> Store OTPs in encrypted form (e.g., AES-256) and expire after 5 minutes (or less).
>> Anti-Phishing Measures :
>> Display partial user information (e.g., last 4 digits of the mobile number or masked name) during OTP entry.
>> 
>> 7. Audit & Monitoring
>> Audit Logs :
>> Log all sensitive actions (e.g., login failures, password resets).
>> Monitoring :
>> Track failed OTP attempts and unusual login patterns with Prometheus/Grafana or ELK Stack .
>> 
>> 8. Third-Party Integrations
>> Secure Webhooks :
>> Sign payloads with HMAC for webhook notifications.
>> External Services :
>> Use mutual TLS (mTLS) for critical integrations (e.g., payment gateways).
>> 
>> 9. Compliance
>> GDPR/CCPA :
>> Allow users to export/delete their data via /self/data-export and /self/delete-account.
>> PCI-DSS :
>> Tokenize payment data and avoid storing raw card details.
>> 
>> 
>> 10. PKI
>> 
>> For making financial transactions the payment payloads can be encrypted/decrypted using X.509, this is the standard that we use in México for Real Time Payments (QR and Push)
>> 
>> 10.a. Strong Client Authentication
>> Mutual TLS (mTLS) :
>> X.509 enables mutual authentication (client and server validate each other), ensuring only trusted devices/users can access payment APIs.
>> Example: A mobile banking app uses a client certificate to authenticate API requests, preventing unauthorized access even if credentials are compromised.
>> 
>> 10.b End-to-End Encryption
>> Data Integrity & Confidentiality :
>> Encrypt API communication using TLS certificates, protecting sensitive payment data (e.g., account numbers, transaction details) from interception or tampering.
>> 
>> 10.c. Non-Repudiation
>> Digital Signatures :
>> Transactions signed with X.509 certificates provide irrefutable proof of origin, critical for dispute resolution in real-time payments.
>> Example: A payment request signed by a client certificate cannot be denied by the sender.
>> 
>> 10.d. Compliance with Financial Regulations
>> PSD2, PCI DSS, and Open Banking :
>> X.509 certificates meet strict regulatory requirements for secure communication and strong customer authentication (SCA).
>> Example: PSD2 mandates certificate-based authentication for payment service providers (PSPs).
>> 
>> 10.e. Granular Access Control
>> Role-Based Permissions :
>> Embed roles (e.g., merchant, user, admin) in certificates to enforce fine-grained authorization.
>> Example: A merchant certificate grants access to /payments but not /admin endpoints.
>> 
>> 10.f. Protection Against MITM Attacks
>> Certificate Pinning :
>> Prevent man-in-the-middle attacks by binding API clients to specific server certificates.
>> Example: Mobile apps pin the server’s X.509 certificate to reject fake certificates.
>> 
>> 10.g. Scalability for High-Volume Transactions
>> Efficient Handshake :
>> TLS session resumption reduces latency for repeated API calls, ideal for real-time payments.
>> Example: A payment gateway processes thousands of transactions/sec with minimal overhead.
>> 
>> 10.h. Automated Certificate Management
>> Short-Lived Certificates :
>> Use tools like Let’s Encrypt or HashiCorp Vault to automate certificate rotation, reducing manual errors.
>> Example: Certificates expire every 90 days and auto-renew without downtime.
>> 
>> 10.i. Integration with Hardware Security Modules (HSMs)
>> Secure Key Storage :
>> Store private keys in HSMs to protect against server-side breaches.
>> Example: Payment APIs use HSM-backed certificates for signing transactions.
>> 
>> 10.j. Auditability
>> Certificate Transparency :
>> Track certificate issuance and usage via logs, enabling forensic analysis of suspicious transactions.
>> Example: Audit logs show which certificate initiated a payment at 2024-03-05T10:00:00Z.
>> Key Considerations about using PKI
>> Certificate Revocation : Use CRL or OCSP to revoke compromised certificates.
>> Performance : Optimize certificate validation to avoid latency in real-time transactions.
>> User Experience : Automate certificate issuance (e.g., via mobile app integration).
>> Key Tools
>> Tool
>> Purpose
>> Keycloak, Zitadel, Gluu or Auth0
>> Identity and access management
>> Spring Security
>> OAuth2/JWT integration
>> OWASP ZAP
>> API vulnerability scanning
>> HashiCorp Vault
>> Secrets management
>> Prometheus/Grafana
>> Monitoring and alerting
>> 
>> By combining these practices, tools, technologies we ensure a secure, user-friendly self-service API architecture but if we focus our efforts on this the innovation on financial products could be slower.
>> 
>> Regards
>> 
>> Victor Romero
>> 
>> P.S 1. Maybe it is time to look at post-quantum cryptography (PQC) algorithms (NIST has some references) for the Apache Fineract itself.
>> P.S. 2. By the way this text is a part of a long detailed architecture for Self Service APIs.
>> P.S. 3. I encourage all of you to start looking at the PRs and Jira ticker and help us with improvements, reading the tickets, code will help to have a good understanding about the project :)
>> P.S. 4. The Apache Fineract source code is as secure as its weakest line of code.
>> P.S. 5. Tools and purpose are broader....
>> 
>> El jue, 13 mar 2025 a las 22:09, James Dailey (<jdailey@apache.org <ma...@apache.org>>) escribió:
>>> All - 
>>> 
>>> To be clear, the design concerns about the existing Self Service component have existed from before it came over from MifosX to Fineract, so this is a long standing problem.  The design suffers from simply being the wrong technical approach, but the concept of "self service" whereby the end-user gains access to the backend for authenticated and authorized activities is - of course - a key functionality.  
>>> 
>>> In the near very near term, our intention at the PMC is to address some security model concerns (being non-specific here) and then to figure out the right kind of component going forward.  We should not rule out a separate component at the Apache Fineract community to handle this functionality.   I think we actually should consider different design options.  
>>> 
>>> Mifos has offered and I think Ed just mentioned it for the first time publicly on this list, to build a "stand in" for the existing functionality that currently sits inside the fineract authorization model.  This will hopefully address some short term concerns.  I will leave it to them to explain their approach else-thread. 
>>> 
>>> In the Apache approach, mifos is a vendor bundling up their solution on top of the fineract project.  I mention this because really any vendor is welcome to attempt what Mifos is doing and offer that to the community of users or to bring a proposal of the same to Fineract.  I would argue that for the reasons articulated around open banking, that we will need to address the architectural elements of this set of functionality in the longer term within the fineract project, but I think people should vote with their feet.  
>>> 
>>> I'll also use this as a plug for getting involved in the project, particularly if you have API security framework or open banking expertise.  Or to be involved in the Fineract Security team.  There's a lot to do on this front.  We're not giving up on "self service" we just need to address it in phases.  
>>> 
>>> Thank you, 
>>> James 
>>> 


Re: Self-Service

Posted by James Dailey <ja...@gmail.com>.
Victor - These are important requirements that you raised.  So far, I think
some people are on board with these things but I don't yet see a strong
effort on this.
What would you advise as the sequence and prioritization?

How does this relate to the ongoing FSIPs?

Do we need to create some tickets to work on?

James Dailey



On Thu, Mar 13, 2025 at 9:19 PM VICTOR MANUEL ROMERO RODRIGUEZ <
victor.romero@fintecheando.mx> wrote:

> Hello,
>
> The Apache Fineract APIs must be deprecated because there are more
> efficient and secure ways to allow end users to access its financial
> services.
>
> The Apache Fineract must continue to include financial products on its
> core modules and it has  to promote the ecosystem around it through
> interfaces that allow third party systems based on open standards the
> interchange of information in a secure manner.
>
> Don't worry there are already plugins (Open sourced which are compatible
> with the current Apache Fineract version)
>
> A self service APi Architecture and Implementation is not limited to the
> technologies that I have mentioned before (On Wed, 31 Jan 2024). Grab a cup
> of coffee and start reading....
>
> For a more detailed explanation about how to design a secure Self-Service
> API REST architecture with access/identity management, OTP, and
> best-in-class security for the Self Service APIs we must design and
> implement:
>
>
> ------------------------------
>
> 1. Authentication & Identity Management
>
>    -
>
>    OAuth 2.0/OpenID Connect :
>    -
>
>       Use Keycloak, Zitadel, Gluu or Auth0 as an identity provider (IdP)
>       for token-based authentication.
>       -
>
>       Example: Secure endpoints with @PreAuthorize("hasRole('USER')") in
>       Spring Boot.
>       -
>
>       Enforce PKCE (Proof Key for Code Exchange) for public clients
>       (mobile apps).
>       -
>
>    JWT Tokens :
>    -
>
>       Use short-lived access tokens (e.g., 5 minutes or less) and refresh
>       tokens.
>       -
>
>       Validate JWT signatures, issuer (iss), audience (aud), and
>       expiration (exp).
>       -
>
>    Password Policies :
>    -
>
>       Enforce strong passwords (min 12 characters, mix of
>       symbols/numbers).
>       -
>
>       Hash passwords with BCrypt or Argon2 (and never ever store it using
>       plaintext).
>
> 2. Multi-Factor Authentication (MFA) with OTP
>
>    -
>
>    TOTP (Time-Based OTP) :
>    -
>
>       Integrate with Aegis, Google, Azure, Authenticator or Twilio Authy.
>       -
>
>       Use Keycloak, Zitadel, Gluu built-in OTP policies for user
>       registration/login.
>       -
>
>    OTP Delivery :
>    -
>
>       Send OTPs via secure channels (e.g., SMS, email, or authenticator
>       apps).
>       -
>
>       Avoid SMS for high-security scenarios (vulnerable to SIM swaps).
>       -
>
>    Rate Limiting :
>    -
>
>       Limit OTP requests (e.g., 3 attempts per hour) to prevent
>       brute-force attacks using an API Gateway (Kong, Gravitee, ApiSix) .
>
> ------------------------------
>
> 3. Authorization & Access Control
>
>    -
>
>    RBAC (Role-Based Access Control) :
>    -
>
>       Define roles like USER, ADMIN, and SUPPORT with granular
>       permissions.
>       -
>
>       Example: Restrict /self/clients/{id} to the owning user only.
>       -
>
>    Scope-Based Access :
>    -
>
>       Use OAuth2 scopes (e.g., read:profile, write:loans) to limit API
>       actions.
>       -
>
>    Attribute-Based Access Control (ABAC) :
>    -
>
>       Restrict access based on user attributes (e.g., country, client_type
>       ).
>
> ------------------------------
>
> 4. Data Protection
>
>    -
>
>    Encryption :
>    -
>
>       In Transit : Enforce HTTPS (TLS 1.3+) using .
>       -
>
>       In Transit: Make sure to use secure CipherSpecs and CipherSuites
>       -
>
>       At Rest : Encrypt sensitive fields (e.g., PII) in the database.
>       -
>
>    Masking Sensitive Data :
>    -
>
>       Log/display masked data (e.g., accountNumber: "CL-****123").
>       -
>
>       Use @JsonView in Spring to control data exposure.
>
> ------------------------------
>
> 5. API Security Best Practices
>
>    -
>
>    Input Validation :
>    -
>
>       Sanitize inputs to prevent SQLi/XSS (e.g., use
>       Hibernate/Apache/Spring Validator annotations).
>       -
>
>    Token Binding :
>    -
>
>       Bind tokens to specific devices/IPs to prevent token theft.
>
> ------------------------------
>
> 6. Secure OTP Workflow
>
>    1.
>
>    Step-Up Authentication :
>    -
>
>       Require OTP for sensitive actions (e.g., changing passwords or
>       transferring funds).
>       2.
>
>    Secure OTP Storage :
>    -
>
>       Store OTPs in encrypted form (e.g., AES-256) and expire after 5
>       minutes (or less).
>       3.
>
>    Anti-Phishing Measures :
>    -
>
>       Display partial user information (e.g., last 4 digits of the mobile
>       number or masked name) during OTP entry.
>
> ------------------------------
>
> 7. Audit & Monitoring
>
>    -
>
>    Audit Logs :
>    -
>
>       Log all sensitive actions (e.g., login failures, password resets).
>       -
>
>    Monitoring :
>    -
>
>       Track failed OTP attempts and unusual login patterns with
>       Prometheus/Grafana or ELK Stack .
>
> ------------------------------
>
> 8. Third-Party Integrations
>
>    -
>
>    Secure Webhooks :
>    -
>
>       Sign payloads with HMAC for webhook notifications.
>       -
>
>    External Services :
>    -
>
>       Use mutual TLS (mTLS) for critical integrations (e.g., payment
>       gateways).
>
> ------------------------------
>
> 9. Compliance
>
>    -
>
>    GDPR/CCPA :
>    -
>
>       Allow users to export/delete their data via /self/data-export and
>       /self/delete-account.
>       -
>
>    PCI-DSS :
>    -
>
>       Tokenize payment data and avoid storing raw card details.
>
>
> ------------------------------
>
> 10. PKI
>
> For making financial transactions the payment payloads can be
> encrypted/decrypted using X.509, this is the standard that we use in México
> for Real Time Payments (QR and Push)
>
> 10.a. Strong Client Authentication
>
>    -
>
>    Mutual TLS (mTLS) :
>    X.509 enables mutual authentication (client and server validate each
>    other), ensuring only trusted devices/users can access payment APIs.
>    -
>
>       Example: A mobile banking app uses a client certificate to
>       authenticate API requests, preventing unauthorized access even if
>       credentials are compromised.
>
> ------------------------------
>
> 10.b End-to-End Encryption
>
>    -
>
>    Data Integrity & Confidentiality :
>    Encrypt API communication using TLS certificates, protecting sensitive
>    payment data (e.g., account numbers, transaction details) from interception
>    or tampering.
>
> ------------------------------
>
> 10.c. Non-Repudiation
>
>    -
>
>    Digital Signatures :
>    Transactions signed with X.509 certificates provide irrefutable proof
>    of origin, critical for dispute resolution in real-time payments.
>    -
>
>       Example: A payment request signed by a client certificate cannot be
>       denied by the sender.
>
> ------------------------------
>
> 10.d. Compliance with Financial Regulations
>
>    -
>
>    PSD2, PCI DSS, and Open Banking :
>    X.509 certificates meet strict regulatory requirements for secure
>    communication and strong customer authentication (SCA).
>    -
>
>       Example: PSD2 mandates certificate-based authentication for payment
>       service providers (PSPs).
>
> ------------------------------
>
> 10.e. Granular Access Control
>
>    -
>
>    Role-Based Permissions :
>    Embed roles (e.g., merchant, user, admin) in certificates to enforce
>    fine-grained authorization.
>    -
>
>       Example: A merchant certificate grants access to /payments but not
>       /admin endpoints.
>
> ------------------------------
>
> 10.f. Protection Against MITM Attacks
>
>    -
>
>    Certificate Pinning :
>    Prevent man-in-the-middle attacks by binding API clients to specific
>    server certificates.
>    -
>
>       Example: Mobile apps pin the server’s X.509 certificate to reject
>       fake certificates.
>
> ------------------------------
>
> 10.g. Scalability for High-Volume Transactions
>
>    -
>
>    Efficient Handshake :
>    TLS session resumption reduces latency for repeated API calls, ideal
>    for real-time payments.
>    -
>
>       Example: A payment gateway processes thousands of transactions/sec
>       with minimal overhead.
>
> ------------------------------
>
> 10.h. Automated Certificate Management
>
>    -
>
>    Short-Lived Certificates :
>    Use tools like Let’s Encrypt or HashiCorp Vault to automate
>    certificate rotation, reducing manual errors.
>    -
>
>       Example: Certificates expire every 90 days and auto-renew without
>       downtime.
>
> ------------------------------
>
> 10.i. Integration with Hardware Security Modules (HSMs)
>
>    -
>
>    Secure Key Storage :
>    Store private keys in HSMs to protect against server-side breaches.
>    -
>
>       Example: Payment APIs use HSM-backed certificates for signing
>       transactions.
>
> ------------------------------
>
> 10.j. Auditability
>
>    -
>
>    Certificate Transparency :
>    Track certificate issuance and usage via logs, enabling forensic
>    analysis of suspicious transactions.
>    -
>
>       Example: Audit logs show which certificate initiated a payment at
>       2024-03-05T10:00:00Z.
>
> Key Considerations about using PKI
>
>    -
>
>    Certificate Revocation : Use CRL or OCSP to revoke compromised
>    certificates.
>    -
>
>    Performance : Optimize certificate validation to avoid latency in
>    real-time transactions.
>    -
>
>    User Experience : Automate certificate issuance (e.g., via mobile app
>    integration).
>
> Key Tools
>
> Tool
>
> Purpose
>
> Keycloak, Zitadel, Gluu or Auth0
>
> Identity and access management
>
> Spring Security
>
> OAuth2/JWT integration
>
> OWASP ZAP
>
> API vulnerability scanning
>
> HashiCorp Vault
>
> Secrets management
>
> Prometheus/Grafana
>
> Monitoring and alerting
>
>
> By combining these practices, tools, technologies we ensure a secure,
> user-friendly self-service API architecture but if we focus our efforts on
> this the innovation on financial products could be slower.
>
>
> Regards
>
> Victor Romero
>
> P.S 1. Maybe it is time to look at post-quantum cryptography (PQC)
> algorithms (NIST has some references) for the Apache Fineract itself.
> P.S. 2. By the way this text is a part of a long detailed architecture for
> Self Service APIs.
> P.S. 3. I encourage all of you to start looking at the PRs and Jira
> ticker and help us with improvements, reading the tickets, code will help
> to have a good understanding about the project :)
> P.S. 4. The Apache Fineract source code is as secure as its weakest line
> of code.
> P.S. 5. Tools and purpose are broader....
>
> El jue, 13 mar 2025 a las 22:09, James Dailey (<jd...@apache.org>)
> escribió:
>
>> All -
>>
>> To be clear, the design concerns about the existing Self Service
>> component have existed from before it came over from MifosX to Fineract, so
>> this is a long standing problem.  The design suffers from simply being the
>> wrong technical approach, but the concept of "self service" whereby the
>> end-user gains access to the backend for authenticated and authorized
>> activities is - of course - a key functionality.
>>
>> In the near very near term, our intention at the PMC is to address some
>> security model concerns (being non-specific here) and then to figure out
>> the right kind of component going forward.  We should not rule out a
>> separate component at the Apache Fineract community to handle this
>> functionality.   I think we actually should consider different design
>> options.
>>
>> Mifos has offered and I think Ed just mentioned it for the first time
>> publicly on this list, to build a "stand in" for the existing functionality
>> that currently sits inside the fineract authorization model.  This will
>> hopefully address some short term concerns.  I will leave it to them to
>> explain their approach else-thread.
>>
>> In the Apache approach, mifos is a vendor bundling up their solution on
>> top of the fineract project.  I mention this because really any vendor is
>> welcome to attempt what Mifos is doing and offer that to the community of
>> users or to bring a proposal of the same to Fineract.  I would argue that
>> for the reasons articulated around open banking, that we will need to
>> address the architectural elements of this set of functionality in the
>> longer term within the fineract project, but I think people should vote
>> with their feet.
>>
>> I'll also use this as a plug for getting involved in the project,
>> particularly if you have API security framework or open banking expertise.
>> Or to be involved in the Fineract Security team.  There's a lot to do on
>> this front.  We're not giving up on "self service" we just need to address
>> it in phases.
>>
>> Thank you,
>> James
>>
>>

Re: Self-Service

Posted by VICTOR MANUEL ROMERO RODRIGUEZ <vi...@fintecheando.mx>.
Hello,

The Apache Fineract APIs must be deprecated because there are more
efficient and secure ways to allow end users to access its financial
services.

The Apache Fineract must continue to include financial products on its core
modules and it has  to promote the ecosystem around it through interfaces
that allow third party systems based on open standards the interchange of
information in a secure manner.

Don't worry there are already plugins (Open sourced which are compatible
with the current Apache Fineract version)

A self service APi Architecture and Implementation is not limited to the
technologies that I have mentioned before (On Wed, 31 Jan 2024). Grab a cup
of coffee and start reading....

For a more detailed explanation about how to design a secure Self-Service
API REST architecture with access/identity management, OTP, and
best-in-class security for the Self Service APIs we must design and
implement:


------------------------------

1. Authentication & Identity Management

   -

   OAuth 2.0/OpenID Connect :
   -

      Use Keycloak, Zitadel, Gluu or Auth0 as an identity provider (IdP)
      for token-based authentication.
      -

      Example: Secure endpoints with @PreAuthorize("hasRole('USER')") in
      Spring Boot.
      -

      Enforce PKCE (Proof Key for Code Exchange) for public clients (mobile
      apps).
      -

   JWT Tokens :
   -

      Use short-lived access tokens (e.g., 5 minutes or less) and refresh
      tokens.
      -

      Validate JWT signatures, issuer (iss), audience (aud), and expiration
      (exp).
      -

   Password Policies :
   -

      Enforce strong passwords (min 12 characters, mix of symbols/numbers).
      -

      Hash passwords with BCrypt or Argon2 (and never ever store it using
      plaintext).

2. Multi-Factor Authentication (MFA) with OTP

   -

   TOTP (Time-Based OTP) :
   -

      Integrate with Aegis, Google, Azure, Authenticator or Twilio Authy.
      -

      Use Keycloak, Zitadel, Gluu built-in OTP policies for user
      registration/login.
      -

   OTP Delivery :
   -

      Send OTPs via secure channels (e.g., SMS, email, or authenticator
      apps).
      -

      Avoid SMS for high-security scenarios (vulnerable to SIM swaps).
      -

   Rate Limiting :
   -

      Limit OTP requests (e.g., 3 attempts per hour) to prevent brute-force
      attacks using an API Gateway (Kong, Gravitee, ApiSix) .

------------------------------

3. Authorization & Access Control

   -

   RBAC (Role-Based Access Control) :
   -

      Define roles like USER, ADMIN, and SUPPORT with granular permissions.
      -

      Example: Restrict /self/clients/{id} to the owning user only.
      -

   Scope-Based Access :
   -

      Use OAuth2 scopes (e.g., read:profile, write:loans) to limit API
      actions.
      -

   Attribute-Based Access Control (ABAC) :
   -

      Restrict access based on user attributes (e.g., country, client_type).

------------------------------

4. Data Protection

   -

   Encryption :
   -

      In Transit : Enforce HTTPS (TLS 1.3+) using .
      -

      In Transit: Make sure to use secure CipherSpecs and CipherSuites
      -

      At Rest : Encrypt sensitive fields (e.g., PII) in the database.
      -

   Masking Sensitive Data :
   -

      Log/display masked data (e.g., accountNumber: "CL-****123").
      -

      Use @JsonView in Spring to control data exposure.

------------------------------

5. API Security Best Practices

   -

   Input Validation :
   -

      Sanitize inputs to prevent SQLi/XSS (e.g., use
      Hibernate/Apache/Spring Validator annotations).
      -

   Token Binding :
   -

      Bind tokens to specific devices/IPs to prevent token theft.

------------------------------

6. Secure OTP Workflow

   1.

   Step-Up Authentication :
   -

      Require OTP for sensitive actions (e.g., changing passwords or
      transferring funds).
      2.

   Secure OTP Storage :
   -

      Store OTPs in encrypted form (e.g., AES-256) and expire after 5
      minutes (or less).
      3.

   Anti-Phishing Measures :
   -

      Display partial user information (e.g., last 4 digits of the mobile
      number or masked name) during OTP entry.

------------------------------

7. Audit & Monitoring

   -

   Audit Logs :
   -

      Log all sensitive actions (e.g., login failures, password resets).
      -

   Monitoring :
   -

      Track failed OTP attempts and unusual login patterns with
      Prometheus/Grafana or ELK Stack .

------------------------------

8. Third-Party Integrations

   -

   Secure Webhooks :
   -

      Sign payloads with HMAC for webhook notifications.
      -

   External Services :
   -

      Use mutual TLS (mTLS) for critical integrations (e.g., payment
      gateways).

------------------------------

9. Compliance

   -

   GDPR/CCPA :
   -

      Allow users to export/delete their data via /self/data-export and
      /self/delete-account.
      -

   PCI-DSS :
   -

      Tokenize payment data and avoid storing raw card details.


------------------------------

10. PKI

For making financial transactions the payment payloads can be
encrypted/decrypted using X.509, this is the standard that we use in México
for Real Time Payments (QR and Push)

10.a. Strong Client Authentication

   -

   Mutual TLS (mTLS) :
   X.509 enables mutual authentication (client and server validate each
   other), ensuring only trusted devices/users can access payment APIs.
   -

      Example: A mobile banking app uses a client certificate to
      authenticate API requests, preventing unauthorized access even if
      credentials are compromised.

------------------------------

10.b End-to-End Encryption

   -

   Data Integrity & Confidentiality :
   Encrypt API communication using TLS certificates, protecting sensitive
   payment data (e.g., account numbers, transaction details) from interception
   or tampering.

------------------------------

10.c. Non-Repudiation

   -

   Digital Signatures :
   Transactions signed with X.509 certificates provide irrefutable proof of
   origin, critical for dispute resolution in real-time payments.
   -

      Example: A payment request signed by a client certificate cannot be
      denied by the sender.

------------------------------

10.d. Compliance with Financial Regulations

   -

   PSD2, PCI DSS, and Open Banking :
   X.509 certificates meet strict regulatory requirements for secure
   communication and strong customer authentication (SCA).
   -

      Example: PSD2 mandates certificate-based authentication for payment
      service providers (PSPs).

------------------------------

10.e. Granular Access Control

   -

   Role-Based Permissions :
   Embed roles (e.g., merchant, user, admin) in certificates to enforce
   fine-grained authorization.
   -

      Example: A merchant certificate grants access to /payments but not
      /admin endpoints.

------------------------------

10.f. Protection Against MITM Attacks

   -

   Certificate Pinning :
   Prevent man-in-the-middle attacks by binding API clients to specific
   server certificates.
   -

      Example: Mobile apps pin the server’s X.509 certificate to reject
      fake certificates.

------------------------------

10.g. Scalability for High-Volume Transactions

   -

   Efficient Handshake :
   TLS session resumption reduces latency for repeated API calls, ideal for
   real-time payments.
   -

      Example: A payment gateway processes thousands of transactions/sec
      with minimal overhead.

------------------------------

10.h. Automated Certificate Management

   -

   Short-Lived Certificates :
   Use tools like Let’s Encrypt or HashiCorp Vault to automate certificate
   rotation, reducing manual errors.
   -

      Example: Certificates expire every 90 days and auto-renew without
      downtime.

------------------------------

10.i. Integration with Hardware Security Modules (HSMs)

   -

   Secure Key Storage :
   Store private keys in HSMs to protect against server-side breaches.
   -

      Example: Payment APIs use HSM-backed certificates for signing
      transactions.

------------------------------

10.j. Auditability

   -

   Certificate Transparency :
   Track certificate issuance and usage via logs, enabling forensic
   analysis of suspicious transactions.
   -

      Example: Audit logs show which certificate initiated a payment at
      2024-03-05T10:00:00Z.

Key Considerations about using PKI

   -

   Certificate Revocation : Use CRL or OCSP to revoke compromised
   certificates.
   -

   Performance : Optimize certificate validation to avoid latency in
   real-time transactions.
   -

   User Experience : Automate certificate issuance (e.g., via mobile app
   integration).

Key Tools

Tool

Purpose

Keycloak, Zitadel, Gluu or Auth0

Identity and access management

Spring Security

OAuth2/JWT integration

OWASP ZAP

API vulnerability scanning

HashiCorp Vault

Secrets management

Prometheus/Grafana

Monitoring and alerting


By combining these practices, tools, technologies we ensure a secure,
user-friendly self-service API architecture but if we focus our efforts on
this the innovation on financial products could be slower.


Regards

Victor Romero

P.S 1. Maybe it is time to look at post-quantum cryptography (PQC)
algorithms (NIST has some references) for the Apache Fineract itself.
P.S. 2. By the way this text is a part of a long detailed architecture for
Self Service APIs.
P.S. 3. I encourage all of you to start looking at the PRs and Jira ticker
and help us with improvements, reading the tickets, code will help to have
a good understanding about the project :)
P.S. 4. The Apache Fineract source code is as secure as its weakest line of
code.
P.S. 5. Tools and purpose are broader....

El jue, 13 mar 2025 a las 22:09, James Dailey (<jd...@apache.org>)
escribió:

> All -
>
> To be clear, the design concerns about the existing Self Service component
> have existed from before it came over from MifosX to Fineract, so this is a
> long standing problem.  The design suffers from simply being the wrong
> technical approach, but the concept of "self service" whereby the end-user
> gains access to the backend for authenticated and authorized activities is
> - of course - a key functionality.
>
> In the near very near term, our intention at the PMC is to address some
> security model concerns (being non-specific here) and then to figure out
> the right kind of component going forward.  We should not rule out a
> separate component at the Apache Fineract community to handle this
> functionality.   I think we actually should consider different design
> options.
>
> Mifos has offered and I think Ed just mentioned it for the first time
> publicly on this list, to build a "stand in" for the existing functionality
> that currently sits inside the fineract authorization model.  This will
> hopefully address some short term concerns.  I will leave it to them to
> explain their approach else-thread.
>
> In the Apache approach, mifos is a vendor bundling up their solution on
> top of the fineract project.  I mention this because really any vendor is
> welcome to attempt what Mifos is doing and offer that to the community of
> users or to bring a proposal of the same to Fineract.  I would argue that
> for the reasons articulated around open banking, that we will need to
> address the architectural elements of this set of functionality in the
> longer term within the fineract project, but I think people should vote
> with their feet.
>
> I'll also use this as a plug for getting involved in the project,
> particularly if you have API security framework or open banking expertise.
> Or to be involved in the Fineract Security team.  There's a lot to do on
> this front.  We're not giving up on "self service" we just need to address
> it in phases.
>
> Thank you,
> James
>
>

Re: Self-Service

Posted by James Dailey <jd...@apache.org>.
All -

To be clear, the design concerns about the existing Self Service component
have existed from before it came over from MifosX to Fineract, so this is a
long standing problem.  The design suffers from simply being the wrong
technical approach, but the concept of "self service" whereby the end-user
gains access to the backend for authenticated and authorized activities is
- of course - a key functionality.

In the near very near term, our intention at the PMC is to address some
security model concerns (being non-specific here) and then to figure out
the right kind of component going forward.  We should not rule out a
separate component at the Apache Fineract community to handle this
functionality.   I think we actually should consider different design
options.

Mifos has offered and I think Ed just mentioned it for the first time
publicly on this list, to build a "stand in" for the existing functionality
that currently sits inside the fineract authorization model.  This will
hopefully address some short term concerns.  I will leave it to them to
explain their approach else-thread.

In the Apache approach, mifos is a vendor bundling up their solution on top
of the fineract project.  I mention this because really any vendor is
welcome to attempt what Mifos is doing and offer that to the community of
users or to bring a proposal of the same to Fineract.  I would argue that
for the reasons articulated around open banking, that we will need to
address the architectural elements of this set of functionality in the
longer term within the fineract project, but I think people should vote
with their feet.

I'll also use this as a plug for getting involved in the project,
particularly if you have API security framework or open banking expertise.
Or to be involved in the Fineract Security team.  There's a lot to do on
this front.  We're not giving up on "self service" we just need to address
it in phases.

Thank you,
James

Re: Self-Service

Posted by Paul <pc...@gmail.com>.
That's great to hear, Ed. Thanks for the insight . . .
I've been getting different answers from Fineract / MIFOS community and
each seems to have their share of uncertainty on this subject self service
features.

On Thu, Mar 13, 2025 at 10:00 PM Ed Cable <ed...@mifos.org> wrote:

> The Mifos Initiative recognizes how critical the self-service capabilities
> are to the users of Fineract and Mifos and do require a more secure way of
> connecting to and administering trusted first and third party customer
> facing applications. Mifos and members of the community are working on the
> design and implementation of this plugin which will be maintained as a
> separate open source Mifos plug-in on top of Fineract and welcome any
> feedback or input into its design.
>
> Thanks,
>
> Ed
>
> On Thu, Mar 13, 2025 at 10:08 PM Paul <pc...@gmail.com> wrote:
>
>> I'm not technical enough to speak about the tech portion, but from a
>> business perspective / opinion, Fineract MUST support self service in their
>> db schema or become intentionally obsolete.  Fineract does not need to
>> support the connector or gateway.  They already exist in multiple flavors.
>>
>> The fundamental regulatory and ethical principles of *Open Banking* emphasize
>> consumer control, which necessitates a self-service function.  The two can
>> not be separated . . .  As of today, 81 countries have already implemented
>> an open data / open banking policy (required by law) with ~17 more
>> implementing / phasing in policies over the next year, (US being one of
>> them). That's 98 countries of 192 chartered in the UN . . . .  have already
>> codified consumer rights to the data and others are following.
>>
>> Regardless of the structure of the connector, node or gateway method
>> which certainly could (and maybe should) be outside the core product, at
>> the end of the day, those gateways must still connect to the Fineract
>> backend and the supporting data points should NOT be custom in my business
>> opinion.
>>
>> Regards, Paul
>>
>>
>> On Thu, Mar 13, 2025 at 8:31 AM Felix van Hove <fv...@gmx.de.invalid>
>> wrote:
>>
>>> Hi all,
>>>
>>> Do you intend to withdraw the capability to *manage* self service users
>>> and potentially certain features of the Self Service app by non-end
>>> users (which is hinted at by menu items in the Mifos X client) too?
>>>
>>> On Wed, 31 Jan 2024, 21:45 VICTOR MANUEL ROMERO RODRIGUEZ, <
>>> victor.rom...@fintecheando.mx> wrote:
>>>  > As an alternative you can use solutions for accessing the APIs like
>>> API
>>>  > Gateways and Access Management.
>>>
>>> Is there any guideline or anything that explains e.g. to a smallish
>>> microcredit org how to set something like this up in the future? I
>>> assume you don't want to put the user management into the API gateway
>>> too.
>>>
>>> Felix
>>>
>>> On 12/03/2025 20:07, James Dailey wrote:
>>> > Yes please.  It would make sense now to remove it - it will still be
>>> > "there" in the code if someone really wants to figure it out but
>>> having it
>>> > gone would reduce some surface area of application concerns.  Also
>>> > vendors have had enough notice on this to have come up with their own
>>> > approach.
>>> >
>>> > Expanding this...
>>> > I also think Fineract should probably build its own "reference
>>> version" for
>>> > the same functionality but in a different github at Apache Fineract.
>>>  This
>>> > would fit better into an overall architecture, where you have the
>>> component
>>> > using secure methods to connect to the backend APIs and then exposing
>>> > different customer facing APIs .  The auth is different for such a
>>> > component, as it relies on identity models for the end user, which are
>>> > fundamentally different than identity validation for employees of a
>>> bank,
>>> > credit union, etc
>>> >
>>> >
>>> >
>>> > James Dailey
>>> >
>>> >
>>> >
>>> > On Wed, Mar 12, 2025 at 8:50 AM Ádám Sághy <ad...@gmail.com>
>>> wrote:
>>> >
>>> >> Hi guys,
>>> >>
>>> >> I would like to pick up once again this topic: Self-service in
>>> Fineract.
>>> >>
>>> >> Since we just had a release and this functionality was still not
>>> removed,
>>> >> i was wondering whether we want to do it now and by the time of the
>>> next
>>> >> release we dont need to worry about that anymore.
>>> >>
>>> >> This was already a topic a couple times whether the existence of this
>>> >> functionality in Fineract is good or bad, but If I recall correctly,
>>> it was
>>> >> majorly considered as a bad thing (problematic from security point of
>>> view
>>> >> as well).
>>> >>
>>> >> Should we have a casual or official vote regarding the self-service
>>> >> functionality?
>>> >>
>>> >> Regards,
>>> >> Adam
>>> >>
>>> >> On 2023. Nov 15., at 14:34, mumin iddrisu <mi...@gmail.com>
>>> wrote:
>>> >>
>>> >> Hello James
>>> >> Thank you for the feedback.
>>> >> I will be looking out for the solution in the coming months.
>>> >>
>>> >> Regards
>>> >> Abdul
>>> >> ---mis
>>> >>
>>> >>
>>> >> On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jd...@apache.org>
>>> wrote:
>>> >>
>>> >>> Abdul -
>>> >>>
>>> >>> We encourage everyone to move away from the self-service component.
>>> We
>>> >>> are going to be revising this shortly as it is not designed for a
>>> >>> production environment.
>>> >>> In lieu of this we will propose a solution.  This is probably a
>>> month or
>>> >>> more away.
>>> >>>
>>> >>> The history is that we've identified issues with this component and
>>> have
>>> >>> slated it, in many roadmap decisions, for immediate removal.  It is
>>> getting
>>> >>> close to that point now; we do need to find a way to describe what
>>> to do
>>> >>> without the Self Service component.
>>> >>>
>>> >>> Jdailey
>>> >>> VP Apache Fineract
>>> >>>
>>> >>> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com>
>>> >>> wrote:
>>> >>>
>>> >>>> Hello Team
>>> >>>> Has anyone successfully deployed the self-service feature?
>>> >>>> If yes please let me know which build you are using.
>>> >>>>
>>> >>>> Regards
>>> >>>> Abdul
>>> >>>> ---mis
>>> >>>> ᐧ
>>> >>>>
>>> >>>
>>> >>
>>> >
>>>
>>>
>>
>> --
>> --
>> Paul
>>
>
>
> --
> *Ed Cable*
> President/CEO, Mifos Initiative
> edcable@mifos.org | Skype: edcable | Mobile: +1.484.477.8649
>
> *Collectively Creating a World of 3 Billion Maries | *http://mifos.org
> <http://facebook.com/mifos>  <http://www.twitter.com/mifos>
>
>

-- 
--
Paul

Re: Self-Service

Posted by Ed Cable <ed...@mifos.org>.
The Mifos Initiative recognizes how critical the self-service capabilities
are to the users of Fineract and Mifos and do require a more secure way of
connecting to and administering trusted first and third party customer
facing applications. Mifos and members of the community are working on the
design and implementation of this plugin which will be maintained as a
separate open source Mifos plug-in on top of Fineract and welcome any
feedback or input into its design.

Thanks,

Ed

On Thu, Mar 13, 2025 at 10:08 PM Paul <pc...@gmail.com> wrote:

> I'm not technical enough to speak about the tech portion, but from a
> business perspective / opinion, Fineract MUST support self service in their
> db schema or become intentionally obsolete.  Fineract does not need to
> support the connector or gateway.  They already exist in multiple flavors.
>
> The fundamental regulatory and ethical principles of *Open Banking* emphasize
> consumer control, which necessitates a self-service function.  The two can
> not be separated . . .  As of today, 81 countries have already implemented
> an open data / open banking policy (required by law) with ~17 more
> implementing / phasing in policies over the next year, (US being one of
> them). That's 98 countries of 192 chartered in the UN . . . .  have already
> codified consumer rights to the data and others are following.
>
> Regardless of the structure of the connector, node or gateway method which
> certainly could (and maybe should) be outside the core product, at the end
> of the day, those gateways must still connect to the Fineract backend and
> the supporting data points should NOT be custom in my business opinion.
>
> Regards, Paul
>
>
> On Thu, Mar 13, 2025 at 8:31 AM Felix van Hove <fv...@gmx.de.invalid>
> wrote:
>
>> Hi all,
>>
>> Do you intend to withdraw the capability to *manage* self service users
>> and potentially certain features of the Self Service app by non-end
>> users (which is hinted at by menu items in the Mifos X client) too?
>>
>> On Wed, 31 Jan 2024, 21:45 VICTOR MANUEL ROMERO RODRIGUEZ, <
>> victor.rom...@fintecheando.mx> wrote:
>>  > As an alternative you can use solutions for accessing the APIs like API
>>  > Gateways and Access Management.
>>
>> Is there any guideline or anything that explains e.g. to a smallish
>> microcredit org how to set something like this up in the future? I
>> assume you don't want to put the user management into the API gateway too.
>>
>> Felix
>>
>> On 12/03/2025 20:07, James Dailey wrote:
>> > Yes please.  It would make sense now to remove it - it will still be
>> > "there" in the code if someone really wants to figure it out but having
>> it
>> > gone would reduce some surface area of application concerns.  Also
>> > vendors have had enough notice on this to have come up with their own
>> > approach.
>> >
>> > Expanding this...
>> > I also think Fineract should probably build its own "reference version"
>> for
>> > the same functionality but in a different github at Apache Fineract.
>>  This
>> > would fit better into an overall architecture, where you have the
>> component
>> > using secure methods to connect to the backend APIs and then exposing
>> > different customer facing APIs .  The auth is different for such a
>> > component, as it relies on identity models for the end user, which are
>> > fundamentally different than identity validation for employees of a
>> bank,
>> > credit union, etc
>> >
>> >
>> >
>> > James Dailey
>> >
>> >
>> >
>> > On Wed, Mar 12, 2025 at 8:50 AM Ádám Sághy <ad...@gmail.com> wrote:
>> >
>> >> Hi guys,
>> >>
>> >> I would like to pick up once again this topic: Self-service in
>> Fineract.
>> >>
>> >> Since we just had a release and this functionality was still not
>> removed,
>> >> i was wondering whether we want to do it now and by the time of the
>> next
>> >> release we dont need to worry about that anymore.
>> >>
>> >> This was already a topic a couple times whether the existence of this
>> >> functionality in Fineract is good or bad, but If I recall correctly,
>> it was
>> >> majorly considered as a bad thing (problematic from security point of
>> view
>> >> as well).
>> >>
>> >> Should we have a casual or official vote regarding the self-service
>> >> functionality?
>> >>
>> >> Regards,
>> >> Adam
>> >>
>> >> On 2023. Nov 15., at 14:34, mumin iddrisu <mi...@gmail.com> wrote:
>> >>
>> >> Hello James
>> >> Thank you for the feedback.
>> >> I will be looking out for the solution in the coming months.
>> >>
>> >> Regards
>> >> Abdul
>> >> ---mis
>> >>
>> >>
>> >> On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jd...@apache.org>
>> wrote:
>> >>
>> >>> Abdul -
>> >>>
>> >>> We encourage everyone to move away from the self-service component.
>> We
>> >>> are going to be revising this shortly as it is not designed for a
>> >>> production environment.
>> >>> In lieu of this we will propose a solution.  This is probably a month
>> or
>> >>> more away.
>> >>>
>> >>> The history is that we've identified issues with this component and
>> have
>> >>> slated it, in many roadmap decisions, for immediate removal.  It is
>> getting
>> >>> close to that point now; we do need to find a way to describe what to
>> do
>> >>> without the Self Service component.
>> >>>
>> >>> Jdailey
>> >>> VP Apache Fineract
>> >>>
>> >>> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com>
>> >>> wrote:
>> >>>
>> >>>> Hello Team
>> >>>> Has anyone successfully deployed the self-service feature?
>> >>>> If yes please let me know which build you are using.
>> >>>>
>> >>>> Regards
>> >>>> Abdul
>> >>>> ---mis
>> >>>> ᐧ
>> >>>>
>> >>>
>> >>
>> >
>>
>>
>
> --
> --
> Paul
>


-- 
*Ed Cable*
President/CEO, Mifos Initiative
edcable@mifos.org | Skype: edcable | Mobile: +1.484.477.8649

*Collectively Creating a World of 3 Billion Maries | *http://mifos.org
<http://facebook.com/mifos>  <http://www.twitter.com/mifos>

Re: Self-Service

Posted by Paul <pc...@gmail.com>.
I'm not technical enough to speak about the tech portion, but from a
business perspective / opinion, Fineract MUST support self service in their
db schema or become intentionally obsolete.  Fineract does not need to
support the connector or gateway.  They already exist in multiple flavors.

The fundamental regulatory and ethical principles of *Open Banking* emphasize
consumer control, which necessitates a self-service function.  The two can
not be separated . . .  As of today, 81 countries have already implemented
an open data / open banking policy (required by law) with ~17 more
implementing / phasing in policies over the next year, (US being one of
them). That's 98 countries of 192 chartered in the UN . . . .  have already
codified consumer rights to the data and others are following.

Regardless of the structure of the connector, node or gateway method which
certainly could (and maybe should) be outside the core product, at the end
of the day, those gateways must still connect to the Fineract backend and
the supporting data points should NOT be custom in my business opinion.

Regards, Paul


On Thu, Mar 13, 2025 at 8:31 AM Felix van Hove <fv...@gmx.de.invalid>
wrote:

> Hi all,
>
> Do you intend to withdraw the capability to *manage* self service users
> and potentially certain features of the Self Service app by non-end
> users (which is hinted at by menu items in the Mifos X client) too?
>
> On Wed, 31 Jan 2024, 21:45 VICTOR MANUEL ROMERO RODRIGUEZ, <
> victor.rom...@fintecheando.mx> wrote:
>  > As an alternative you can use solutions for accessing the APIs like API
>  > Gateways and Access Management.
>
> Is there any guideline or anything that explains e.g. to a smallish
> microcredit org how to set something like this up in the future? I
> assume you don't want to put the user management into the API gateway too.
>
> Felix
>
> On 12/03/2025 20:07, James Dailey wrote:
> > Yes please.  It would make sense now to remove it - it will still be
> > "there" in the code if someone really wants to figure it out but having
> it
> > gone would reduce some surface area of application concerns.  Also
> > vendors have had enough notice on this to have come up with their own
> > approach.
> >
> > Expanding this...
> > I also think Fineract should probably build its own "reference version"
> for
> > the same functionality but in a different github at Apache Fineract.
>  This
> > would fit better into an overall architecture, where you have the
> component
> > using secure methods to connect to the backend APIs and then exposing
> > different customer facing APIs .  The auth is different for such a
> > component, as it relies on identity models for the end user, which are
> > fundamentally different than identity validation for employees of a bank,
> > credit union, etc
> >
> >
> >
> > James Dailey
> >
> >
> >
> > On Wed, Mar 12, 2025 at 8:50 AM Ádám Sághy <ad...@gmail.com> wrote:
> >
> >> Hi guys,
> >>
> >> I would like to pick up once again this topic: Self-service in Fineract.
> >>
> >> Since we just had a release and this functionality was still not
> removed,
> >> i was wondering whether we want to do it now and by the time of the next
> >> release we dont need to worry about that anymore.
> >>
> >> This was already a topic a couple times whether the existence of this
> >> functionality in Fineract is good or bad, but If I recall correctly, it
> was
> >> majorly considered as a bad thing (problematic from security point of
> view
> >> as well).
> >>
> >> Should we have a casual or official vote regarding the self-service
> >> functionality?
> >>
> >> Regards,
> >> Adam
> >>
> >> On 2023. Nov 15., at 14:34, mumin iddrisu <mi...@gmail.com> wrote:
> >>
> >> Hello James
> >> Thank you for the feedback.
> >> I will be looking out for the solution in the coming months.
> >>
> >> Regards
> >> Abdul
> >> ---mis
> >>
> >>
> >> On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jd...@apache.org>
> wrote:
> >>
> >>> Abdul -
> >>>
> >>> We encourage everyone to move away from the self-service component.  We
> >>> are going to be revising this shortly as it is not designed for a
> >>> production environment.
> >>> In lieu of this we will propose a solution.  This is probably a month
> or
> >>> more away.
> >>>
> >>> The history is that we've identified issues with this component and
> have
> >>> slated it, in many roadmap decisions, for immediate removal.  It is
> getting
> >>> close to that point now; we do need to find a way to describe what to
> do
> >>> without the Self Service component.
> >>>
> >>> Jdailey
> >>> VP Apache Fineract
> >>>
> >>> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com>
> >>> wrote:
> >>>
> >>>> Hello Team
> >>>> Has anyone successfully deployed the self-service feature?
> >>>> If yes please let me know which build you are using.
> >>>>
> >>>> Regards
> >>>> Abdul
> >>>> ---mis
> >>>> ᐧ
> >>>>
> >>>
> >>
> >
>
>

-- 
--
Paul

Re: Self-Service

Posted by Felix van Hove <fv...@gmx.de.INVALID>.
Hi all,

Do you intend to withdraw the capability to *manage* self service users
and potentially certain features of the Self Service app by non-end
users (which is hinted at by menu items in the Mifos X client) too?

On Wed, 31 Jan 2024, 21:45 VICTOR MANUEL ROMERO RODRIGUEZ, <
victor.rom...@fintecheando.mx> wrote:
 > As an alternative you can use solutions for accessing the APIs like API
 > Gateways and Access Management.

Is there any guideline or anything that explains e.g. to a smallish
microcredit org how to set something like this up in the future? I
assume you don't want to put the user management into the API gateway too.

Felix

On 12/03/2025 20:07, James Dailey wrote:
> Yes please.  It would make sense now to remove it - it will still be
> "there" in the code if someone really wants to figure it out but having it
> gone would reduce some surface area of application concerns.  Also
> vendors have had enough notice on this to have come up with their own
> approach.
>
> Expanding this...
> I also think Fineract should probably build its own "reference version" for
> the same functionality but in a different github at Apache Fineract.   This
> would fit better into an overall architecture, where you have the component
> using secure methods to connect to the backend APIs and then exposing
> different customer facing APIs .  The auth is different for such a
> component, as it relies on identity models for the end user, which are
> fundamentally different than identity validation for employees of a bank,
> credit union, etc
>
>
>
> James Dailey
>
>
>
> On Wed, Mar 12, 2025 at 8:50 AM Ádám Sághy <ad...@gmail.com> wrote:
>
>> Hi guys,
>>
>> I would like to pick up once again this topic: Self-service in Fineract.
>>
>> Since we just had a release and this functionality was still not removed,
>> i was wondering whether we want to do it now and by the time of the next
>> release we dont need to worry about that anymore.
>>
>> This was already a topic a couple times whether the existence of this
>> functionality in Fineract is good or bad, but If I recall correctly, it was
>> majorly considered as a bad thing (problematic from security point of view
>> as well).
>>
>> Should we have a casual or official vote regarding the self-service
>> functionality?
>>
>> Regards,
>> Adam
>>
>> On 2023. Nov 15., at 14:34, mumin iddrisu <mi...@gmail.com> wrote:
>>
>> Hello James
>> Thank you for the feedback.
>> I will be looking out for the solution in the coming months.
>>
>> Regards
>> Abdul
>> ---mis
>>
>>
>> On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jd...@apache.org> wrote:
>>
>>> Abdul -
>>>
>>> We encourage everyone to move away from the self-service component.  We
>>> are going to be revising this shortly as it is not designed for a
>>> production environment.
>>> In lieu of this we will propose a solution.  This is probably a month or
>>> more away.
>>>
>>> The history is that we've identified issues with this component and have
>>> slated it, in many roadmap decisions, for immediate removal.  It is getting
>>> close to that point now; we do need to find a way to describe what to do
>>> without the Self Service component.
>>>
>>> Jdailey
>>> VP Apache Fineract
>>>
>>> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com>
>>> wrote:
>>>
>>>> Hello Team
>>>> Has anyone successfully deployed the self-service feature?
>>>> If yes please let me know which build you are using.
>>>>
>>>> Regards
>>>> Abdul
>>>> ---mis
>>>> ᐧ
>>>>
>>>
>>
>


Re: Self-Service

Posted by James Dailey <ja...@gmail.com>.
Yes please.  It would make sense now to remove it - it will still be
"there" in the code if someone really wants to figure it out but having it
gone would reduce some surface area of application concerns.  Also
vendors have had enough notice on this to have come up with their own
approach.

Expanding this...
I also think Fineract should probably build its own "reference version" for
the same functionality but in a different github at Apache Fineract.   This
would fit better into an overall architecture, where you have the component
using secure methods to connect to the backend APIs and then exposing
different customer facing APIs .  The auth is different for such a
component, as it relies on identity models for the end user, which are
fundamentally different than identity validation for employees of a bank,
credit union, etc



James Dailey



On Wed, Mar 12, 2025 at 8:50 AM Ádám Sághy <ad...@gmail.com> wrote:

> Hi guys,
>
> I would like to pick up once again this topic: Self-service in Fineract.
>
> Since we just had a release and this functionality was still not removed,
> i was wondering whether we want to do it now and by the time of the next
> release we dont need to worry about that anymore.
>
> This was already a topic a couple times whether the existence of this
> functionality in Fineract is good or bad, but If I recall correctly, it was
> majorly considered as a bad thing (problematic from security point of view
> as well).
>
> Should we have a casual or official vote regarding the self-service
> functionality?
>
> Regards,
> Adam
>
> On 2023. Nov 15., at 14:34, mumin iddrisu <mi...@gmail.com> wrote:
>
> Hello James
> Thank you for the feedback.
> I will be looking out for the solution in the coming months.
>
> Regards
> Abdul
> ---mis
>
>
> On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jd...@apache.org> wrote:
>
>> Abdul -
>>
>> We encourage everyone to move away from the self-service component.  We
>> are going to be revising this shortly as it is not designed for a
>> production environment.
>> In lieu of this we will propose a solution.  This is probably a month or
>> more away.
>>
>> The history is that we've identified issues with this component and have
>> slated it, in many roadmap decisions, for immediate removal.  It is getting
>> close to that point now; we do need to find a way to describe what to do
>> without the Self Service component.
>>
>> Jdailey
>> VP Apache Fineract
>>
>> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com>
>> wrote:
>>
>>> Hello Team
>>> Has anyone successfully deployed the self-service feature?
>>> If yes please let me know which build you are using.
>>>
>>> Regards
>>> Abdul
>>> ---mis
>>> ᐧ
>>>
>>
>

Re: Self-Service

Posted by Ádám Sághy <ad...@gmail.com>.
Hi guys,

I would like to pick up once again this topic: Self-service in Fineract.

Since we just had a release and this functionality was still not removed, i was wondering whether we want to do it now and by the time of the next release we dont need to worry about that anymore.

This was already a topic a couple times whether the existence of this functionality in Fineract is good or bad, but If I recall correctly, it was majorly considered as a bad thing (problematic from security point of view as well).

Should we have a casual or official vote regarding the self-service functionality?

Regards,
Adam

> On 2023. Nov 15., at 14:34, mumin iddrisu <mi...@gmail.com> wrote:
> 
> Hello James
> Thank you for the feedback.
> I will be looking out for the solution in the coming months.
> 
> Regards
> Abdul
> ---mis
> 
> 
> On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jdailey@apache.org <ma...@apache.org>> wrote:
>> Abdul - 
>> 
>> We encourage everyone to move away from the self-service component.  We are going to be revising this shortly as it is not designed for a production environment. 
>> In lieu of this we will propose a solution.  This is probably a month or more away. 
>> 
>> The history is that we've identified issues with this component and have slated it, in many roadmap decisions, for immediate removal.  It is getting close to that point now; we do need to find a way to describe what to do without the Self Service component.  
>> 
>> Jdailey
>> VP Apache Fineract 
>> 
>> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mismis.gh@gmail.com <ma...@gmail.com>> wrote:
>>> Hello Team
>>> Has anyone successfully deployed the self-service feature?
>>> If yes please let me know which build you are using.
>>> 
>>> Regards
>>> Abdul
>>> ---mis
>>> ᐧ


Re: Self-Service

Posted by mumin iddrisu <mi...@gmail.com>.
Hello James
Thank you for the feedback.
I will be looking out for the solution in the coming months.

Regards
Abdul
---mis


On Wed, Nov 15, 2023 at 4:01 AM James Dailey <jd...@apache.org> wrote:

> Abdul -
>
> We encourage everyone to move away from the self-service component.  We
> are going to be revising this shortly as it is not designed for a
> production environment.
> In lieu of this we will propose a solution.  This is probably a month or
> more away.
>
> The history is that we've identified issues with this component and have
> slated it, in many roadmap decisions, for immediate removal.  It is getting
> close to that point now; we do need to find a way to describe what to do
> without the Self Service component.
>
> Jdailey
> VP Apache Fineract
>
> On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com> wrote:
>
>> Hello Team
>> Has anyone successfully deployed the self-service feature?
>> If yes please let me know which build you are using.
>>
>> Regards
>> Abdul
>> ---mis
>> ᐧ
>>
>

Re: Self-Service

Posted by James Dailey <jd...@apache.org>.
Abdul -

We encourage everyone to move away from the self-service component.  We are
going to be revising this shortly as it is not designed for a production
environment.
In lieu of this we will propose a solution.  This is probably a month or
more away.

The history is that we've identified issues with this component and have
slated it, in many roadmap decisions, for immediate removal.  It is getting
close to that point now; we do need to find a way to describe what to do
without the Self Service component.

Jdailey
VP Apache Fineract

On Tue, Nov 14, 2023 at 8:43 AM mumin iddrisu <mi...@gmail.com> wrote:

> Hello Team
> Has anyone successfully deployed the self-service feature?
> If yes please let me know which build you are using.
>
> Regards
> Abdul
> ---mis
> ᐧ
>