Storage

Storage

Outrun's storage architecture is designed for maximum reliability, performance, and compliance. Our multi-region infrastructure ensures your data is always available while meeting local data sovereignty requirements.

Multi-Region Architecture

High Availability Design

Every piece of data is stored with built-in redundancy:

• Primary Dataset: Your main operational data
• 2 Replicas: Automatic replication across logically separate data centers
• 3 Data Centers Minimum: Always distributed across at least 3 facilities
• Real-time Synchronization: Changes replicated immediately across all locations

Regional Data Centers

Intelligent Regional Placement

Automatic User-Based Placement

Rather than storing all data in a single region, Outrun intelligently places data based on user location:

// Example: Multi-region data placement
{
  "workspace": "acme-corp",
  "users": [
    {
      "email": "[email protected]",
      "location": "Germany",
      "dataRegion": "EU",
      "primaryDC": "Nuremberg",
      "replicas": ["Falkenstein", "Helsinki"]
    },
    {
      "email": "[email protected]", 
      "location": "Australia",
      "dataRegion": "AU",
      "primaryDC": "S1-Sydney",
      "replicas": ["M2-Melbourne", "P1-Perth"]
    }
  ]
}

Benefits of Regional Placement

• Compliance: Automatic adherence to local data sovereignty laws
• Performance: Reduced latency for users in their home region
• Single Interface: One unified platform regardless of data location
• Regulatory Flexibility: Meet multiple jurisdictions without separate instances

Location Detection Methods

Outrun determines user regions through multiple data sources:

1. CRM Location Fields: User-provided location data in your systems
2. Email Domain Analysis: Geographic indicators from email addresses
3. IP Geolocation: Fallback method for user session data
4. Manual Override: Explicit region selection when needed

Infrastructure Management

Containerized Applications

Our applications are fully containerized for maximum flexibility:

• Multi-Cloud Deployment: Workloads distributed across providers
• Load Balancing: Automatic traffic distribution
• Failover Capability: Seamless switching between providers
• Scaling: Dynamic resource allocation based on demand

Provider Independence

We avoid hyperscaler lock-in for strategic reasons:

Why We Avoid Hyperscalers

• Cost Efficiency: Better value from specialized providers
• Technology Freedom: No vendor lock-in prevents innovation constraints
• Competitive Pricing: Leverage competition between providers
• Service Quality: Focus on providers who excel in specific regions

Regional Provider Strategy

• Local Expertise: Partners domiciled in each region
• Regulatory Alignment: Providers familiar with local compliance requirements
• Cultural Understanding: Better support for regional business practices
• Economic Benefits: Supporting local technology ecosystems

Infrastructure Flexibility

Provider Migration Capability

We maintain the ability to move between providers when necessary:

• Infrastructure Outages: Rapid recovery from provider issues
• Service Improvements: Migration to better performing providers
• Risk Reduction: Diversification across multiple infrastructure partners
• Cost Optimization: Taking advantage of competitive pricing

Migration Process

graph TD
    A[Monitor Provider Performance] --> B{Issue Detected?}
    B -->|Yes| C[Evaluate Alternative Providers]
    B -->|No| A
    C --> D[Plan Migration Strategy]
    D --> E[Execute Gradual Migration]
    E --> F[Validate Data Integrity]
    F --> G[Complete Cutover]
    G --> H[Monitor New Environment]

Data Storage Architecture

Logical Database Separation

Each workspace maintains complete data isolation through separate tenant databases:

• Workspace ID: Unique identifier for data segregation
• Tenant Databases: Separate database schema per workspace (workspace_{workspaceId})
• Access Controls: Strict permissions based on workspace membership
• Data Boundaries: No cross-workspace data access

Storage Layers

Stream Storage

Raw data from sources stored with full fidelity in the stream_data table:

stream_data
├── source_id      → Links to the source that produced this data
├── external_id    → Original record ID from the source system
├── record (JSONB) → Original API response preserved in full
├── metadata (JSONB) → Ingestion timestamps and processing status
└── created_at     → When the record was ingested

Consolidated Storage

Processed and merged data ready for standardization in the consolidated_data table:

consolidated_data
├── source_id       → Links to the originating source
├── external_id     → Unique record identifier (unique per source)
├── object_type     → Classification (people, organizations, etc.)
├── record (JSONB)  → Deduplicated and merged record data
├── metadata (JSONB) → Quality scores, merge history, validation results
└── created_at / updated_at

Standardized Storage

Final standardized objects stored in dedicated typed tables:

people             → Contacts, leads, users
organizations      → Companies, accounts, business entities
relationships      → Connections between people and organizations
search_analytics_data → Search metrics, rankings, and performance data

Performance Optimization

Regional Performance Benefits

• Reduced Latency: Data stored close to users
• Faster Sync: Shorter distances for data replication
• Local Processing: Compute resources in the same region as data
• Bandwidth Efficiency: Minimized cross-region data transfer

Caching Strategy

• Edge Caching: Frequently accessed data cached regionally
• Query Optimization: Intelligent query routing to nearest replica
• Connection Pooling: Efficient database connection management
• Compression: Data compression for network efficiency

Monitoring & Alerting

• Real-time Metrics: Continuous performance monitoring
• Automated Alerts: Proactive notification of issues
• Capacity Planning: Predictive scaling based on usage patterns
• Health Checks: Regular validation of data center connectivity

Disaster Recovery

Business Continuity Planning

• RTO (Recovery Time Objective): < 15 minutes for critical services
• RPO (Recovery Point Objective): < 5 minutes data loss maximum
• Automated Failover: No manual intervention required
• Geographic Distribution: Protection against regional disasters

Backup Strategy

• Continuous Replication: Real-time data synchronization
• Point-in-Time Recovery: Restore to any moment in the last 30 days
• Cross-Region Backups: Additional backups in alternate regions
• Integrity Verification: Regular backup validation and testing

Compliance Benefits

Data Sovereignty

• Local Storage: Data remains within chosen jurisdiction
• Regional Providers: Infrastructure owned by local entities
• Regulatory Alignment: Compliance with local data protection laws
• Audit Trail: Complete record of data location and movement

Multi-Jurisdiction Support

• GDPR Compliance: EU data stored within EU boundaries
• Australian Privacy Act: Local data protection compliance

Future Expansion

Planned Regions

We're continuously expanding our global footprint:

• Canada: Toronto and Vancouver data centers
• United Kingdom: London and Manchester facilities
• Singapore: Southeast Asia regional hub
• Japan: Tokyo metropolitan area coverage

Enhanced Capabilities

• Edge Computing: Processing closer to data sources
• AI/ML Integration: Regional machine learning capabilities
• Real-time Analytics: Low-latency data processing
• IoT Support: Internet of Things data ingestion

Best Practices

Region Selection

1. Assess User Distribution: Understand where your users are located
2. Review Compliance Requirements: Consider regulatory obligations
3. Evaluate Performance Needs: Balance latency vs. compliance
4. Plan for Growth: Consider future expansion plans

Data Management

1. Monitor Regional Distribution: Track data placement across regions
2. Review Access Patterns: Optimize based on usage patterns
3. Validate Compliance: Regular compliance audits
4. Plan Capacity: Anticipate storage growth needs

Next Steps

Outrun's storage architecture ensures your data is always available, compliant, and performant - no matter where your users are located.