top of page

Your Comprehensive Guide to Parallel Data Warehouse

Introduction to Parallel Data Warehouse

Parallel Data Warehouse (PDW), part of the Analytics Platform System, represents a significant advancement in data warehousing and analytics. Designed to handle large-scale data processing, PDW employs a massively parallel processing (MPP) architecture that distributes data and queries across multiple nodes. This architecture enables rapid query performance, efficient data movement, and robust data management capabilities. In this comprehensive guide, we will explore the core components of PDW, its functionalities, and how it can revolutionize data warehousing for businesses.

Understanding Parallel Data Warehouse Components

PDW comprises several key components, each playing a critical role in the overall functionality of the system. These components include the Control Node, Compute Nodes, Data Movement Service (DMS), and various supporting software and hardware elements.

Appliance Software - Query Processing and User Data Storage

Control Node

The Control Node is the central coordinator in the PDW system. It manages query processing, user data storage, and overall system operations. Key functions include:

  • MPP Engine: The MPP Engine is the brain of the PDW system. It creates parallel query plans, coordinates parallel query execution across Compute Nodes, and manages metadata and configuration data for all databases. It also handles SQL Server PDW database authentication and authorization, and tracks hardware and software status.

  • Data Movement Service (DMS): DMS is critical for data transfer operations. It moves data between SQL Server PDW nodes, optimizes data transfer speeds, and enhances query performance by efficiently processing data that requires movement among nodes.

  • Admin Console: This web application provides a comprehensive view of the appliance state, health, and performance. It is accessible over HTTPS and allows administrators to monitor and manage the system.

  • Configuration Manager: The Configuration Manager (dwconfig.exe) is used by appliance administrators to configure various aspects of the Analytics Platform System.

Compute Nodes

Compute Nodes are the workhorses of the PDW system. They handle parallel data processing and storage, and each node has its own direct-attached storage managed by SQL Server. Key functions include:

Compute nodes

  • Data Movement Service (DMS): Similar to the Control Node, DMS on Compute Nodes transfers data to and from other Compute Nodes and the Control Node. It also handles data loading in parallel, directly from the loading server to the Compute Nodes, and manages data transfer for backups and integration with external systems like Hadoop or Azure Storage Blob.

  • Compute Node Databases: Each Compute Node runs an instance of SQL Server to process queries and manage user data.

Appliance Fabric

The appliance fabric provides the underlying infrastructure for the PDW system, including the operating system, services, and network infrastructure.

Domain Controller and Active Directory

Active Directory (AD) Domain Services manage authentication across the PDW nodes and handle SQL Server PDW Windows Authentication logins.

DNS Service

The Windows Domain Name Service (DNS) resolves domain names to IP addresses within the PDW appliance.

Windows Deployment Service

Windows Deployment Service (WDS) deploys the Windows Server operating system onto the appliance. It ensures that all hosts and virtual machines across the appliance are correctly configured and running the necessary software.

Virtual Machine Manager

The Virtual Machine Manager hosts System Center, deploying the operating system on physical hosts. It also manages Windows Server Update Services (WSUS) for applying or removing updates across all hosts and virtual machines.

Failover Clustering and Storage Spaces

Windows Failover Clustering ensures high availability by restarting processes on a passive host if a failure occurs. Windows Storage Spaces manage user data as a storage pool for Compute Nodes, ensuring data accessibility even if a node fails.


Microsoft Hyper-V Server provides a reliable virtualization solution, balancing CPU resources and ensuring high availability for PDW nodes and appliance fabric components.

Handling Non-Relational Data with PolyBase

PolyBase technology integrates SQL Server PDW data with external Hadoop data. This integration allows PDW to query and manage data stored in Hadoop clusters or Azure Storage Blob seamlessly. Supported Hadoop distributions include Hortonworks, Cloudera, and HDInsight.

Query Tools and Business Intelligence Integration

PDW supports various tools and interfaces for querying and integrating data, enhancing its utility for business intelligence and analytics.

Query Tools

  • SQL Server Data Tools (SSDT): Running inside Visual Studio, SSDT is the recommended GUI tool for submitting queries to SQL Server PDW. It offers an intuitive interface similar to SQL Server Management Studio.

  • sqlcmd Command-Line Query Tool: sqlcmd is the command-line tool for running Transact-SQL statements and system commands. It supports interactive query execution, batch files, and integration with Windows PowerShell.

  • Integration Services: SQL Server Integration Services (SSIS) can be used to query and load data into PDW.

  • Linked Server: A SQL Server linked server connection allows using SQL Server to submit Transact-SQL statements to PDW.

Business Intelligence Tools

  • Analysis Services: PDW serves as a data source for Analysis Services databases and Excel PowerPivot models, supporting both multidimensional and relational online analytical processing.

  • Report Builder: PDW can be used as a SQL Server data source for reports developed with SQL Server Report Builder.

  • Power Pivot for Excel: PowerPivot for Excel significantly expands Excel’s data analysis capabilities by connecting to PDW.

Loading Tools and Data Integration

PDW supports various tools and methods for loading data, ensuring efficient and flexible data integration.

Integration Services and dwloader

  • Integration Services: PDW-specific destination adapters allow using SSIS to load data into PDW.

  • dwloader Command-Line Loader: dwloader is a command-line tool that loads data in parallel from your loading server to PDW Compute Nodes.

PolyBase for Hadoop Integration

PolyBase enables loading non-relational data from Hadoop clusters into PDW relational tables, integrating data stored externally with the PDW system.

Database Backup and Restore

PDW uses Transact-SQL database backup and restore commands to handle database backups and restores in parallel, ensuring data integrity and recovery capabilities. Backups are written to a Windows file share, from which they can also be restored.

Remote Table Copy

The Remote Table Copy feature allows copying tables from PDW databases to remote SQL Server databases, enabling hub-and-spoke data distribution scenarios.

Monitoring and Management

PDW provides multiple ways to monitor appliance activity and manage the system efficiently.

Admin Console

The Admin Console is a web application that displays the current status and health of the appliance. It is accessible over HTTPS.

System Views

System views provide detailed information about the appliance’s status and performance. Administrators can query these views to obtain specific data.

System Center Operations Manager

System Center Operations Manager (SCOM) Management Packs for PDW enable comprehensive monitoring and management of the appliance.


Parallel Data Warehouse (PDW) is a robust solution for large-scale data processing and analytics, offering high performance, scalability, and flexibility. By leveraging its MPP architecture, PDW can efficiently manage massive datasets, integrate with various data sources, and support complex queries and business intelligence operations. Understanding its components and functionalities is crucial for maximizing its potential and ensuring optimal performance in your data warehousing environment.

Key Takeaways

  1. MPP Architecture: PDW utilizes a massively parallel processing architecture to distribute data and queries across Compute Nodes, enhancing query performance and scalability.

  2. Core Components: It consists of Control Nodes for management and Compute Nodes for processing data, supported by services like Data Movement Service (DMS) for efficient data transfers.

  3. Integration with Hadoop: PolyBase technology enables seamless integration with Hadoop clusters and Azure Storage Blob, allowing PDW to query and manage non-relational data alongside relational data.

  4. Query and BI Tools: PDW supports SQL Server Data Tools (SSDT), sqlcmd for command-line queries, and integration with SQL Server Integration Services (SSIS) for data loading and transformation.

  5. Data Loading and Integration: Various tools like dwloader and PolyBase facilitate efficient data loading from external sources into PDW, maintaining data integrity and scalability.

  6. Backup and Restore: PDW uses Transact-SQL commands for parallel database backups and restores, ensuring robust data protection and recovery capabilities.

  7. Monitoring and Management: Admin Console, system views, and System Center Operations Manager (SCOM) provide comprehensive monitoring and management capabilities for PDW appliances.

  8. Scalability and Performance: Designed for large-scale environments, PDW offers high performance, scalability, and flexibility for complex data warehousing and analytics needs.


What is a Parallel Data Warehouse (PDW)?

PDW is part of the Analytics Platform System, designed for large-scale data processing using a massively parallel processing (MPP) architecture.

How does PDW handle data queries?

PDW uses the MPP Engine to create and execute parallel query plans across Compute Nodes, optimizing performance and efficiency.

What is the role of the Data Movement Service (DMS) in PDW?

DMS transfers data between nodes, processes queries that require data movement, and optimizes data transfer speeds.

How does PDW integrate with Hadoop?

PolyBase technology allows PDW to query and manage data stored in Hadoop clusters or Azure Storage Blob, supporting various Hadoop distributions.

What tools are available for querying PDW?

Tools include SQL Server Data Tools (SSDT), sqlcmd command-line tool, Integration Services, Linked Server, and various business intelligence tools like Analysis Services and Report Builder.

How can I monitor and manage a PDW appliance?

PDW provides monitoring tools like the Admin Console, system views, and System Center Operations Manager (SCOM) Management Packs.

Article Sources


bottom of page