Skip to main content

Payment Orchestration

Azotte's payment orchestration engine intelligently routes payments across multiple providers for optimal success rates and cost efficiency.

Orchestration Overview​

┌─────────────┐     ┌─────────────────────┐     ┌──────────────┐
│  Customer   │────▶│  Payment Gateway    │────▶│ PSP Selector │
│  Checkout   │     │                     │     │              │
└─────────────┘     │ • Tokenization      │     │ • Routing    │
                    │ • Validation        │     │ • Failover   │
                    │ • Fraud Detection   │     │ • Retry      │
                    └─────────┬───────────┘     └──────┬───────┘
                              │                         │
                    ┌─────────▼──────────────────┬──────▼───────┐
                    │      Payment Providers      │              │
                    │                            │              │
            ┌───────▼──┐ ┌────────▼──┐ ┌────────▼──┐ ┌────────▼──┐
            │  Stripe  │ │   Adyen   │ │ Craftgate │ │  PayPal   │
            │          │ │           │ │           │ │           │
            └──────────┘ └───────────┘ └───────────┘ └───────────┘

Routing Intelligence​

Routing Factors​

  • Geographic Optimization: Region-specific PSP performance
  • Payment Method: Method-specific routing preferences
  • Success Rate: Historical performance data
  • Cost Optimization: Transaction fee minimization
  • Currency Support: Native currency processing
  • Feature Requirements: Advanced features needed

Smart Routing Algorithm​

1. Analyze Transaction
   ├─ Customer Location
   ├─ Payment Method
   ├─ Transaction Amount
   └─ Currency

2. Score Available PSPs
   ├─ Success Rate Score (40%)
   ├─ Cost Score (30%)
   ├─ Performance Score (20%)
   └─ Feature Score (10%)

3. Select Primary & Backup PSPs
   ├─ Primary: Highest scoring PSP
   └─ Backup: Next 2-3 highest scoring

4. Execute with Failover
   ├─ Attempt Primary PSP
   ├─ Auto-retry on soft failures
   └─ Cascade to backup PSPs

Failover and Retry Logic​

Failure Types​

  • Soft Failures: Temporary issues, retry-able

    • Network timeouts
    • Rate limiting
    • Temporary PSP downtime

  • Hard Failures: Permanent issues, no retry

    • Invalid credentials
    • Insufficient funds
    • Fraud detection blocks

Retry Strategy​

  • Exponential backoff with jitter
  • Maximum retry limits per PSP
  • Circuit breaker pattern implementation
  • Intelligent retry scheduling

Cascade Failover​

  1. Primary PSP failure detection
  2. Automatic secondary PSP selection
  3. Transaction context preservation
  4. Seamless customer experience

PSP Adapter Pattern​

Standardized Interface​

interface PaymentProvider {
  charge(request: ChargeRequest): Promise<ChargeResponse>
  refund(request: RefundRequest): Promise<RefundResponse>
  tokenize(request: TokenRequest): Promise<TokenResponse>
  webhook(payload: WebhookPayload): WebhookEvent
}

Provider-Specific Adapters​

  • Stripe Adapter: Stripe-specific implementation
  • Adyen Adapter: Adyen-specific implementation
  • PayPal Adapter: PayPal-specific implementation
  • Generic Adapter: Common PSP interface

Real-Time Monitoring​

Performance Metrics​

  • Success rate per PSP
  • Average response time
  • Error rate analysis
  • Cost per transaction

Health Monitoring​

  • PSP availability monitoring
  • Response time tracking
  • Error pattern detection
  • Capacity utilization

Alerting System​

  • PSP downtime notifications
  • Success rate degradation alerts
  • Cost threshold warnings
  • Performance anomaly detection

Configuration Management​

PSP Configuration​

  • Connection parameters
  • Feature capability mapping
  • Geographic availability
  • Cost structure definitions

Routing Rules​

  • Priority-based routing
  • Condition-based routing
  • Time-based routing
  • Volume-based routing

Security and Compliance​

PCI DSS Compliance​

  • Tokenization at edge
  • Secure key management
  • Encrypted data transmission
  • Regular security audits

Fraud Prevention​

  • Multi-layer fraud detection
  • Machine learning models
  • Real-time risk scoring
  • Behavioral analysis

Next Steps​