Data Mesh and Data Fabric

The data mesh and data fabric are the latest incarnations of data architectures to gain credence in our industry. The first three—operational reporting, data warehousing, and data lakehouses—are all largely centralized, IT-directed environments. In contrast, data mesh and data fabric are inherently business driven and decentralized—so a major break from the past. (See figure 1.) 

Each data architecture had its limitations which propelled the creation and adoption of a new data architecture. The question is what will be the limitations of the data mesh and data fabric? 

Customer Adoption. Although there is a lot of buzz around data fabric and data mesh architectures, there is little adoption yet. Only 7.3% of Chief Data & Analytics Officers (CDAOs) indicate either is a top priority for 2023 although 41.5% say they are investing in it. A much larger percentage of CDAOs (80%) are investing in data products. (See figure 2.) 

If we separate the two architectures, we see that Gartner is bullish on data fabric, as is Wayne, who believes that every company needs a data fabric to supplement a central data environment and inject agility into it. As for data mesh, a majority of customers are trying to learn about data mesh or waiting to see how early adopters fare. Like data products, there are many ideas about what a data mesh is and how to implement it. This has caused many companies to hold back on investing time and resources into building a data mesh.

“A data fabric injects agility into a monolithic data environment—so it’s a vital supplement to data warehouses and lakehouses.”

“There is a lot of confusion about what a data mesh is, which is holding back adoption.”

Figure 1. Evolution of Data Architectures


Figure 2. Market Attitudes toward Data Mesh and Data Fabric

View Large