GigaOm Sonar Report for Data Fabric Solutionsv1.01

1. Summary

A data fabric is an emerging approach to data architecture, comprehensively connecting all of an organization’s data sources and data processing within a unified layer in which everything is accessible through a single interface. A data fabric provides this functionality regardless of differences in the physical locations of native sources, which span hybrid cloud, multicloud, and on-premises environments. Data fabrics also standardize data to overcome differences in formatting, schema, data models, nomenclature, and file types.

Providing this universal access to all data sources integrated into a standardized format is immensely valuable to end users in many use cases and scenarios. The data fabric abstracts the complexity of the data integration and standardization process, so organizations simply access data from their tool(s) of choice with minimal knowledge of those technical details. Common consuming clients include teams handling business intelligence (BI), data science, and streaming data, along with developers and operational applications.

This architecture simultaneously solves the problem of the increasing distribution of the data landscape, in which data is scattered about on seemingly endless external and internal systems, and of accessing a broad range of data from a multitude of sources. It enables organizations to access all data from everywhere uniformly, as though it came from in-house sources, and in a format that’s readily understood in easy, business-friendly terms.

Implementation types include physical data fabrics, in which data is moved into a common ecosystem like that of some of the hyperscalers, and logical data fabrics. The latter involves technologies such as data virtualization, materialization, and query federation, so data never moves away from sources—or does so at the last moment possible. A data fabric incorporates all aspects of data management (data preparation, data engineering, data modeling, data governance, master data management, and data security) so that data products may be easily consumed.

The data fabric concept also involves some of the most progressive notions of the data sphere. Organizations are realizing this architecture is complementary, not an alternative, to the data mesh architecture. Data mesh operates as a bottom-up approach to building reusable data products via individual business units with central oversight for data governance by IT teams. A data fabric provides a top-down stance in which organizations can combine data products (supplied by a data mesh) for use cases such as analytics across marketing, product, and sales data. Advanced solutions combine artificial intelligence and metadata to automate the data integrations that are foundational to data fabric utility. Semantic inferencing and supervised and unsupervised learning techniques support this development to provide scalable automation for data fabric deployments.

This GigaOm Sonar provides an overview of data fabric vendors and their available offerings, equipping IT decision-makers with the information they need to select the best solution for their business and use case requirements.