Deploying Mojaloop
This section details the deployment aspects of the Mojaloop Hub.
Performance Baseline
The Mojaloop Hub has been demonstrated to support the following performance characteristics on minimal hardware:
Clearing 1,000 transfers per second
Sustained for one hour
With not more than 1% (of transfer stage) taking more than 1 second through the hub
This baseline performance can be used as a reference point for system sizing and capacity planning.
Deployment of Mojaloop Hub (excluding participant integrations)
The following table provides guidance on which Mojaloop deployment scenario is most appropriate for different user types and use cases.
Student
Footprint: Single machine e.g. laptop or single VM. SLA: None
?
?
Footprint: Single machine e.g. laptop or single VM. SLA: None
N/A
Developer
Footprint: Single machine e.g. laptop or single VM. SLA: None
?
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
N/A
Business analyst
Footprint: Single machine e.g. laptop or single VM. SLA: None
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
N/A
Potential adopter
Footprint: Single machine e.g. laptop or single VM. SLA: None
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
N/A
Auditor / External QA / Security Analyst
Footprint: Single machine e.g. laptop or single VM. SLA: None
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
N/A
N/A
System Integrator
Footprint: Single machine e.g. laptop or single VM. SLA: None
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Fully redundant, replicated, high availability deployment - On-premises or cloud SLA: High SLA in many areas.
Hub Operator
Footprint: Single machine e.g. laptop or single VM. SLA: None
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Single machine e.g. laptop or single VM. - Production like environment (sandbox? lower SLA than prod) SLA: - Lower than prod but possibility of testing NFRs.
Footprint: - Fully redundant, replicated, high availability deployment - On-premises or cloud SLA: High SLA in many areas.
Deployment Tools
core test harness
- single node - docker-compose - "profiles" available - No HELM - No gateway - No ingress/egress components - No IAM stack - Deploys: - core services & backing services - portals (optional) - monitoring stack (optional) Used in CI pipelines for integration testing.
Mid level laptop or desktop workstation
Zero security
- Developer focused documentation. - Non-technical user documentation to support learning objectives. - Product level documentation to explain features e.g. what it does and where it is appropriate for use
- No SLA
- Never to be used in production. - Never to be used to process real money transactions.
Miniloop
- single node kubernetes - microk8s? - HELM with string replacements - No gateway - No ingress/egress components - No IAM stack - Deploys: - core services & backing services Gives ability to test in Kubernetes layer.
Mid level laptop or desktop workstation
Zero security
- Developer focused documentation - Semi-technical / BA focused documentation to support use-case experimentation and testing. - Product level documentation to explain features e.g. what it does and where it is appropriate for use.
- No SLA
- Never to be used in production. - Never to be used to process real money transactions. - Possibly useful for testing/understanding of the HELM charts.
HELM deploy
- Just HELM charts needed to deploy Mojaloop services and backing services.
- High end laptop or workstation - Small cloud kubernetes cluster.
User is required to harden their own Kubernetes cluster.
- Developer focused documentation - Semi-technical / BA focused documentation to support use-case experimentation and testing. - Product level documentation to explain features e.g. what it does and where it is appropriate for use.
- Must be able to achieve SLAs (given baseline hardware specs): - Availability: - ? 4/5 9's? - RTO/RPO: ? As close to zero as possible. - Throughput/Performance - TPS: 1000+ (sustained for 1 hour) - Latency (Percentiles) (excluding external latencies): - Clearing: 99% < 1 second. - Lookup: 99% < second. - Agreement of Terms: 99% < 1 second. - Data management: - Mitigations against data loss i.e. replication, disaster recovery. - Retention (audit, compliance) - Archiving. NB: Strategy is high availability over disaster recovery.
- Can be used in production. - Safe for processing real money transactions. - User/adopter is required to deploy and configure their own infrastructure, including Kubernetes cluster(s), ingress/egress, firewalls etc... - Security is limited to what HELM charts provide. Additional security design and configuration is required.
IaC
- multiple deployment platform targets - AWS, On-prem, other clouds, (modular) - multiple orchestration layer options - managed k8s, microk8s, eks - GitOps pattern (control centre) - can deploy and manage multiple hub instances / environments - Deploys: - control centre - core services & backing services (options for managed backing services) - portals - IAM stack - monitoring stack - pm4ml - GitOps pattern
- High end cloud or on-premise infrastructure.
Full security
- Multiple levels of documentation targeting all levels of "user". - Developer docs to enable use, maintenance, enhancement, extenstion of IaC capabilities e.g. adding new targets / services / features. - Detailed architecture diagrams and explanation to enable deep understanding. - Tech ops focused docs to enable "infrastructure engineer" level users to use IaC to deploy and maintain multiple mojaloop instances for development, testing and production. - Product level documentation to explain IaC features e.g. what it does and where it is appropriate for use.
- Must be able to achieve SLAs (given baseline hardware specs): - Availability: - ? 4/5 9's? - RTO/RPO: ? As close to zero as possible. - Throughput/Performance - TPS: 1000+ (sustained for 1 hour) - Latency (Percentiles) (excluding external latencies): - Clearing: 99% < 1 second. - Lookup: 99% < second. - Agreement of Terms: 99% < 1 second. - Data management: - Mitigations against data loss i.e. replication, disaster recovery. - Retention (audit, compliance) - Archiving. NB: Strategy is high availability over disaster recovery.
- Can be used in production. - Safe for processing real money transactions.
Document History
1.0
7th May 2025
Tony Williams
Initial version
Last updated
Was this helpful?