Overview
- Contoso, Ltd. is a company that sells enriched financial data to a variety of external customers.
- Contoso has a main office in Los Angeles and two branch offices in New York and Seattle.
Existing Environment
Data Infrastructure
- Contoso has a 50-TB data warehouse that uses an instance of SQL Server on Azure Virtual Machines.
- The data warehouse populates an Azure Synapse Analytics workspace that is accessed by the external customers. Currently, the customers can access all the data.
- Contoso has one Power BI workspace named FinData that contains a single dataset. The dataset contains financial data from around the world. The workspace is used by 10 internal users and one external customer. The dataset has the following two data sources: the data warehouse and the Synapse Analytics serverless SQL pool.
- Users frequently query the Synapse Analytics workspace by using Transact-SQL.
User Problems
Contoso identifies the following user issues:
- Some users indicate that the visuals in Power BI reports are slow to render when making filter selections.
- Users indicate that queries against the serverless SQL pool fail occasionally because the size of tempdb has been exceeded.
- Users indicate that the data in Power BI reports is stale. You discover that the refresh process of the Power BI model occasionally times out.
Planned Changes
Contoso plans to implement the following changes:
- Into the existing Power BI dataset, integrate an external data source in JSON that is accessible by using the REST API.
- Build a new dataset in the FinData workspace by using data from the Synapse Analytics dedicated SQL pool.
- Provide all the customers with their own Power BI workspace to create their own reports. Each workspace will use the new dataset in the FinData workspace.
- Implement subscription levels for the customers. Each subscription level will provide access to specific rows of financial data.
- Deploy prebuilt datasets to Power BI to simplify the query experience of the customers.
- Provide internal users with the ability to incorporate machine learning models loaded to the dedicated SQL pool.