Search docs.molo17.com
    A newer version of this documentation is available.
    View Latest (v2.3)

    Gluesync Agents 101: Functional Architecture Overview

    Gluesync’s architecture is built on a powerful multi-layered system of specialized agents, each handling specific aspects of the data integration process.

    Core Components

    Gluesync’s functional architecture diagram

    CDC Engine (Source Connector)

    Purpose: Manages the "first mile" connection between data sources and the ingestion engine.

    Key Features:

    Feature Description

    Security

    * Native TLS support * Vendor-certified JDBC drivers * Official SDK implementations

    Connectivity

    * JDBC for relational databases * Native SDKs for NoSQL/data lakes * Auto-recovery for connection losses

    Consistency

    Part of Gluesync’s two-phase commit system

    We exclusively use vendor-issued drivers and SDKs - no custom implementations.

    Data Streaming Engine (Target Connector)

    Purpose: Handles the "last mile" delivery of data to target systems.

    Key Features:

    Feature Description

    Performance

    * Load balancing across nodes * Optimized write buffers * Multi-node cluster support

    Reliability

    * Auto-recovery mechanisms * Connection loss handling * Transaction consistency

    The streaming engine optimizes performance while maintaining data integrity.

    Data Ingestion Engine

    Purpose: Real-time data processing and transaction management.

    Key Responsibilities:

    Task Description

    Load Distribution

    Multi-threaded processing of data feeds

    State Management

    Maintains the "Gluesync state preservation table"

    Mode Selection

    Determines between full snapshot or change capture

    The engine focuses on data flow without introspecting content, ensuring efficient processing.

    Data Modeling Engine

    Purpose: Real-time data transformation and mapping.

    Key Features:

    Feature Description

    In-Memory Processing

    All transformations happen in RAM

    Schema Management

    Maps source structures to target requirements

    Data Conversion

    Handles format transformations while preserving data integrity

    • Maintains data security by not introspecting content

    • Preserves encrypted and binary data integrity

    • Functions like a real-time ETL processor

    Security Considerations

    • All connections use vendor-certified security implementations

    • Data integrity is maintained throughout the pipeline

    • Encryption is preserved end-to-end

    For specific driver versions and SDK details, contact our support team.