First, many thanks to the SQL Saturday and MNPass team for putting on another great event and letting me participate.
I co-presented with Joshuha Owen (@JoshSQL) in a precon on Modern Enterprise Data Warehousing on Azure. Thanks to those who attended and participated in the conversations around changing the way we implement data warehouse capabilities in Azure. Josh and I will be talking more about this in the future.
Now, those of you who attended my Saturday presentation on Consumption Based Architecture, I wanted to get you the slide deck and reference materials here. Thanks again for attending.
The slide deck from the session is here.
I also have blog post around this topic: Consumption Based Architecture for Modern Data Analytics. Feel free to join the conversation there around this.
ERPs and the Consumption Based Architecture Conversation
One of the key topics that came up during the session was related to handling ERPs with minimal change. The key issue surrounding ERP solutions is with the data structure in those systems. Whether you work with SAP, JackHenry, or Dynamics, you have a situation where the data model is very complex and definitely not user friendly. In Consumption Based Architecture, we try to minimize data transformation and reshaping, but ERP solutions are by nature cryptic and complex. By definition, they are not consumable. So in the consumable space, we typically recommend using the vendor supplied solutions such as JHKnow, SAP BW and so on. These solutions provide a vendor managed interpretation of the data in the ERP for reporting and other solutions.
Security in this Architecture
The question was raised during the session around how to secure this. This does not have a simple answer. Each solution may have implemented security differently. For instance, an Oracle database may use user names and not have AD integration. This means that you need to determine how to secure your consumable space. For instance, if you pick Azure Active Directory, you would move data to AAD compliant structures in Azure such as Azure SQL Database, Azure SQL Datawarehouse, and Azure Databricks. This means you might need to use a tool like Goldengate with CDC to update a SQL DB which you can apply security to. This will allow you to centralize security for your consumable data. You will need to plan for security in whatever you do and create what you need to support it.
Thanks again everyone for joining us at SQL Saturday.