January 17, 2022
With Statista projecting that enterprise data will grow an average of 42.2% over a short two years, it’s no wonder becoming a data-driven organization has become a key priority and challenge among companies of all industries and sizes.
In the session Demystifying Big Data Fabrics and Why You Need One at Hitachi Social Innovation Forum 2021, Hitachi Vantara’s Madhup Mishra met with Forrester Research’s VP and Principal Analyst Noel Yuhanna to discuss how companies can capture value from their data to generate scalable success.
In the last decade, organizations have increasingly struggled with effectively managing their growing data, minimizing silos, and gathering actionable insights. When integrated properly, data fabric has the power to enable businesses to access to their data and address those challenges. Ideally, organizations would like to enable smarter decision-making by being able to:
Data fabric is a modern data management architecture that intelligently orchestrates disparate data sources. “Intelligent” is essential in this context. A data fabric doesn’t just bring together the data sources; it is a framework for doing so in a contextual, secure, integrated manner that enables self-service and supports a wide variety of applications, operational workloads, analytics, and use cases.
The data management layer facilitates the end-to-end integration of the data fabric. From the moment data is captured to the moment it is consumed, this layer is where an organization gets its guarantee of quality. This layer provides a comprehensive, consistent, and uniform framework of data governance and integration that underpins everything else that happens across the data environment. Other layers are answers to essential questions, such as: How do we orchestrate? How do we integrate? How do we transform data? How do we prepare this data? How do we model data?
A benefit to this layered approach is that an organization doesn’t have to deploy everything at once. Instead, it is possible to deploy a few bits of time to address a narrow set of requirements and expand over time. The architecture is low or even no code by design, so new data sources can effectively be dragged and dropped into the framework. The intelligence within the system automatically identifies incoming data, knows which data needs to be transformed, and is aware of what data is suitable to support specific workloads.
Such a system might once have required tools from a hundred vendors, each operating independently of the other. Those separations slowed the entire system as each process had to wait for others to complete. Data fabrics are integrated to provide faster time to value.
But this is not a rip-and-replace scenario. According to Yuhanna, organizations “can leverage existing data management tools. Maybe there’s a catalog. Maybe there is some sort of an ETL already. You have some ETL capabilities. You may already have some streaming capabilities built into your infrastructure.”
The beauty of the data fabric approach is that an organization can grow into it.
One of the threads that bind the data fabric together is semantic enrichment that informs the modeling of days. Knowing the data needed for different business areas and workloads enables the data fabric to drive fast results and create significant value. Semantics enables the data fabric to make sense of data to further the quality, privacy and governance goals of the business.
Yuhanna points out that with data fabric, you don’t have to throw away existing architectures and tools. You can actually complement it. He concludes that organizations can “incrementally grow the fabric, supplementing the existing architecture and driving results.”
Check out more great stories on Insights.