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.