How to Design an Enterprise DWH with this Architecture for On-Premises or Cloud?

aNumak & Company
5 min readApr 7, 2023


A Data Warehouse (DWH) is essential to modern-day business operations. It helps organizations store, process, and analyze vast amounts of data, enabling them to make data-driven decisions. DWH design requires careful consideration of several factors, including architecture, scalability, data integration, security, and performance. This article will guide you through designing an enterprise DWH using the latest architecture for on-premises or cloud deployment.

Understanding the Architecture:

The architecture of a DWH is critical to its performance, scalability, and efficiency. The traditional approach to DWH design involves building a multi-tiered architecture with separate layers for data storage, ETL (Extract, Transform, Load), and reporting. However, the latest trend is to adopt a cloud-based architecture that leverages the benefits of cloud computing, such as scalability, elasticity, and cost-effectiveness.

In the cloud-based architecture, the DWH is deployed on a cloud platform, and the data is stored in a cloud-based data storage service, such as Amazon S3 or Azure Blob Storage. In addition, the ETL process uses cloud-based data processing services like AWS Glue or Azure Data Factory. Finally, the reporting layer is built on the DWH using cloud-based tools like Power BI or Tableau.

Critical Considerations for Designing an Enterprise DWH:

1)Data Integration: Data integration is one of the most critical factors in designing an enterprise DWH. Integrating data means bringing together disparate datasets and converting them into a usable, standardized format for analysis.

The DWH should support integrating structured and unstructured data from multiple sources, such as databases, cloud applications, and social media platforms.

2)Scalability: Scalability is another crucial consideration in DWH design. The DWH should be designed to handle large volumes of data and support growing users. Therefore, the architecture should be flexible enough to allow for easy scalability of resources, such as storage and processing power.

3)Performance: Performance is another critical consideration in DWH design. The DWH should be optimized for fast data retrieval and processing. In addition, the architecture should support complex queries and analytics with minimal latency.

4)Security: Organizations worry a lot about the security of their data, so the DWH should be built with data security in mind.

The DWH should be secured with robust access control measures, encryption, and monitoring.

5)Reporting and Analytics: Reporting and analytics are essential components of a DWH, and the architecture should support a wide range of reporting and analytics tools. The DWH should be designed to provide users with self-service access to data, enabling them to create reports and perform analytics without IT support.

6)Maintenance and Support: Maintenance and support are essential for the long-term success of a DWH. The architecture should be designed to minimize downtime and provide easy access to support services. The DWH should also be prepared to accommodate future updates and upgrades.

Designing for On-Premises:

While cloud-based DWH architecture is gaining popularity, many organizations still prefer to keep it on-premises for regulatory or security reasons. Therefore, designing an enterprise DWH for on-premises deployment requires a different approach than cloud-based deployment.

The traditional multi-tiered architecture is still relevant for on-premises DWH design. The data storage layer can be built using a relational database like SQL Server or Oracle. The ETL process can be executed using tools like Informatica or Talend. Finally, the reporting layer can be built using BI tools like Power BI or Tableau.

Designing for Cloud:

Designing an enterprise DWH for cloud deployment requires careful consideration of several factors, including scalability, security, and cost-effectiveness.

Scalability: Cloud-based DWH architecture allows for easy scalability of resources, making it easier to handle large volumes of data. The DWH can be scaled up or down based on the organization’s requirements, ensuring optimal performance at all times.

Security: Cloud-based DWH architecture requires robust security measures to protect sensitive data from unauthorized access. Data encryption, access control, and monitoring are essential components of a secure cloud-based DWH.

Cost-effectiveness: Cloud-based DWH architecture can be more cost-effective than on-premises deployment, eliminating the need for expensive hardware and software. Organizations can pay for only the resources they need, reducing overall costs.

Steps Involved in Designing an Enterprise DWH:

1)Define Requirements: The first step in designing an enterprise DWH is to define the requirements. This involves understanding the business objectives, the types of data to be collected, the frequency of data collection, and the types of analyses that will be performed. Requirements should be documented clearly and concisely, aligning with the organization’s business strategy.

2)Develop the Data Model: The second step is to develop the data model. The data model should be designed to support the requirements identified in the first step. For example, the data model should include tables, fields, relationships, and constraints and be prepared to optimize data retrieval and processing.

3)Choose the Technology: The third step is to choose the technology. The preferred technology should be based on the organization’s requirements, budget, and IT infrastructure. In addition, the technology should be scalable, secure, and able to handle large volumes of data.

4)Design the Architecture: The fourth step is to design the architecture. The architecture should be prepared to support the data model and technology chosen in the previous actions. In addition, the architecture should be optimized for data integration, scalability, performance, security, reporting, and analytics.

5)Develop ETL Processes: The fifth step is to develop the ETL (Extract, Transform, Load) processes. ETL processes extract data from various sources, transform it into an easily analyzed format, and load it into the DWH. Therefore, the ETL processes should be designed to optimize data integration and processing.

6)Test the System: The sixth step is to test the system. The DWH needs to be tested to make sure it works as intended and according to the organization’s requirements.

Testing should include data validation, performance testing, and security testing.

7)Deploy the System: The final step is to deploy the system. Deployment should be phased, ensuring that the system is fully operational before it is deployed to production. In addition, deployment should include data migration, user training, and support services.


Designing an enterprise DWH requires careful consideration of several factors, including architecture, scalability, security, and cost-effectiveness. Cloud-based DWH architecture is gaining popularity due to its scalability, elasticity, and cost-effectiveness. However, on-premises deployment is still relevant for organizations that prefer to keep their DWH in-house for regulatory or security reasons. The traditional multi-tiered architecture is still suitable for on-premises DWH design, while the cloud-based architecture leverages the benefits of cloud computing. By considering these factors, organizations can design an enterprise DWH that meets their unique requirements.



aNumak & Company

aNumak & Company is a Global Business and Management Consulting firm with expertise in building scalable business models for diverse industry verticals.