Mojaloop Community Central

James Bush
James Bush

Posted on

 

Participation Tools Workstream Weekly Update 2024-10-03

Notes from Participation Tools Worksteam Weekly Sync Call

  • RPi demo - 10tps sequential easily achieved with very basic sequential test from TTK with mock DFSP backend.
    • This is addressing smaller DFSP use cases
  • Payment Manager adoption; there is good stuff there e.g. MCM, management API, onboarding UX etc…:
    • We need a plan to close the gaps on docs + test coverage.
  • Range of solutions to cater to different “classes” of DFSP:
    • Todo: Define these classes properly:
    • Do they self host their CBS? Or use SaaS?
    • Do they have IT skills to install and configure?
    • Small:
      • Option1: low cost “server” hardware e.g. RPi or mini-itx etc…, on-prem, docker-compose+systemd, linux OS firewall
        • Minimal redundancy requirement, can accept some downtime.
        • Todo: figure out UI requirements and how to deliver them efficiently; pm4ml UI? Or something else?
        • Pros: one off cost
        • Cons: needs installation and setup.
        • Could be PISP only via OpenBanking SDK.
        • No need for CBS integration.
      • Option2: aggregator model
        • Do they have their own CBS, bank accounts etc…?
        • ITK hosted by an aggregator?
        • Pros: simple
        • Cons: recurring costs?
        • Possibly suits “micro” DFSPs.
        • Could be PISP only via OpenBanking SDK.
        • No need for CBS integration.
    • Mid-range:
      • Option1: mid-level “server” hardware Some redundancy requirements, less willing to accept downtime.
        • Todo: Find out typical SLAs.
        • More “enterprise” type features for management and monitoring.
        • Possibly integration into existing IT management stack (metrics, observability, alterting)
        • Reporting features?...probably same across the board.
      • Option 2: 3rd party hosting of integration layer Cloud or on-prem
        • Typically requires VPN between integration layer and CBS/backends. Not “mojaloopy”.
        • VPN seen as a cost barrier for small DFSPs in Mojaloop principles….is it ok for larger DFSPs?
          • This is ripe for challenging!
        • Pros: simple
        • Cons: recurring cost
    • Large:
      • Option1: high-end “server” hardware (possibly existing on-prem private/public/hybrid cloud)
        • High resilience, reliability, redundancy requirements. Strong need for enterprise grade management, monitoring, alerting features.
        • High security bar.
        • Scalable performance requirement.
      • Option2: 3rd party hosting of integration layer Cloud or on-prem
        • Strong need for enterprise grade management, monitoring, alerting features.
        • High security bar.
        • Scalable performance requirement.
    • Common:
      • Fast iteration to make use of scheme enhancements (upgrades, more traffic etc…):
      • CI for integration dev work: change -> test -> release
    • Common requirements:
      • “Simple” integration with backends:
        • Low-code?
        • Efficient testing and validation not just during initial setup but ongoing.
          • Dev, UAT, Production testing? CI?
      • Excellent security.
      • Low cost.
      • Efficient use of resources.
      • Ability to “see” what has happened (debugging, reporting)
      • Support???!!!

Please find all meeting notes in this doc: https://docs.google.com/document/d/1FCwkin-2JE4OwjVk9wb3xAjs6aDywgxTdGbg4tJJYsQ/edit?usp=sharing

Top comments (0)