Unified Semantic Layers, Unifying Business and IT
In this post I want to talk about a simple but powerful concept in business intelligence and analytics called a “unified semantic layer.”
You might be a business user or analyst who works with a variety of tools, like Tableau or Qlik, or you might be someone in IT who supports these tools, making them available to business users.
With multiple tools, you are both going to have your challenges. Qlik might be used in Marketing, Tableau in Sales, and each department is going to be limited to the data provided by the data sources that are connected to them. Each tool is going to provide a partial view of the available enterprise information for discovery, reports, and dashboards. Each tool will only provide data to the extent that a particular data source contributes that data for the specific function that the user is performing with that specific tool.
If you are a business user, and you need to get a holistic view of the information, very often you will have to move from tool to tool to assemble this information. And sometimes when you produce a report, it might not tally with another report, even when both reports start with the same information.
If you’re in IT, and you have to support these different tools, you have a related problem. Each of these tools has its own semantic layer, since the data in its raw format, in tables and columns, needs to be translated into business definitions so the reports can be understood by business users. So not only do you have to set up these semantic functions, but you have to do so multiple times – once for Tableau, once for Qlik, once for Business Objects, etc. As a result of all this effort, you might start imposing restrictions on which tool can be used, or the number of tools that are allowed.
Enter the Unified Semantic Layer
Now imagine a single data layer, one that brings together the data from all the different sources within the enterprise. These sources can be within on-premises data centers, they could be in the cloud, or they could come from third-party vendors. This layer would be like a “virtual data lake,” if you will, one that is comprised of all the data from all the different sources. You can then plug in as many BI tools on top of it as you want, to get all the information you need. Now, imagine that this single data layer also contains a single semantic layer that contains all of the semantic definitions of all of the underlying sources, and converts these individual definitions into a common set of business definitions. This is a unified semantic layer.
Let’s look at some of the benefits of a unified semantic layer. First, if you are a business user, you can use any tool to gain a holistic view of all the data across the enterprise. So you can do your discovery, you can produce your reports, you can build your dashboards, everything from your tool of your choice, without having to switch from one tool to another.
If you’re in IT, you can set up a single semantic layer without having to duplicate this function across the different tools. As a result, you can democratize the availability of these tools for your business users.
How to Establish a Unified Semantic Layer
A unified semantic layer can be built in a straightforward manner, using a technology called data virtualization. Data virtualization brings data together from multiple different sources in a virtual, or logical fashion, making the data available to consuming tools in real time. When you request a particular type of data, data virtualization goes to the sources, gets the information, and delivers it in your needed format. Business users get their needed information much more quickly and the data is always fresh. Because data virtualization is implemented as a unified data-access layer above a company’s different data sources, IT stakeholders can define global definitions only once, using fewer resources; they do not have to replicate these changes for every unique tool.
Make Life Simple
The challenge was a siloed view of the information and the difficulty of supporting all the different tools. But with a unified semantic layer supported by data virtualization, business users can gain access to all the enterprise data, in real time. They can discover their data, run reports on it, perform whatever functions they want across the entire data set, and they can make their results available with any business tool. A unified semantic layer makes life simple for both business and IT.
- Data Virtualization and Data Science - July 1, 2021
- Key Insights from Three Cloud Experts Roundtables: Accelerate Hybrid Cloud Journey, Harness Cloud Best Practices, and Simplify Data Management - September 30, 2020
- A CIO’s Guide: How to survive tough economic times through IT Portfolio Rationalization - September 9, 2020