Click to learn more about author Kendall Clark.

The mandate for IT to deliver business value has never been stronger. In fact, 76% of executives believe IT must be an active partner in developing business strategy. Agility is key to success here. However, most enterprises are hampered by data strategies that leave teams flat-footed when the market shifts or new challenges arise.

Take structured Data Management systems, for instance. This option worked well when the enterprise data landscape was itself predominantly structured. But the world is different now, and the enterprise data landscape is now dominated by hybrid, varied, and changing data. The emergence of the Internet of Things (IoT), rise in unstructured data volume, increasing relevance of external data sources, and the trend towards hybrid multi-cloud environments are obstacles to satisfying each new data request. The old Data Strategy, centered around relational data systems, is fundamentally broken. So how can enterprises shift from a reactive to a responsive Data Strategy?

Enterprise Data Fabrics: The Path Forward

Organizations today are looking to build a data fabric to power collaborative, cross-functional projects and products and to escape reactive workflows with a resilient digital foundation – no rip-and-replace required. Data fabrics weave together data from internal data silos and external sources and create a network of information to power apps, AI, and analytics. Quite simply, they support the full breadth of data challenges in today’s complex, connected enterprise.

Unlike older, static data integration techniques, the key principles of data fabrics are that they can:

• Answer unanticipated questions and adapt to new requirements
• Bring meaning to data, which leads to better insight
• Enable queries across data silos and external sources, regardless of data structure
• Modernize existing systems so that no rip-and-replace is required
• Connect data at the compute layer, not at the storage layer, so that data silos can be connected without creating additional silos

Data fabrics also support the cross-functional data connections that are key to creating and defending competitive advantage and enabling collaboration across the enterprise and with external partners. Take as an example the challenges around supply chain innovation. Conventional supply chain data systems are a relay race, operating with linear handoffs and siloed, peer-to-peer links between systems. We saw the predictable results when COVID-19 hit, and global supply chains collapsed. Some strain or even partial collapse was inevitable, but consequences were made worse by inadequate data strategies that treated the supply chain as a rigid system. In reality, the supply chain is a complex network of actors that have to be fully in sync to adjust as needed.

With a digital supply network powered by a data fabric, enterprises can answer complex questions they were previously blind to, such as “show me all the lots of raw materials and associated suppliers involved in the production of finished good lot 123.” Or “how do COGS for product A compare between these two regions?” Or “which manufacturers supplied the raw ingredients involved in this customer complaint?”

Stitching Together a Successful Data Fabric Starts by Understanding Its Materials

Unlike other approaches, data fabrics weave together existing Data Management systems and applications. So, it’s no wonder that data fabrics are quickly being seen as the next step forward in the maturation of the data integration space. This is happening because data fabrics can:

1. Uncover Hidden Meaning: Data fabrics change the status quo by delivering meaning, not just data, across the enterprise. This meaning is woven together from many sources: data and metadata, internal and external sources, and cloud and on-prem systems. Meaning is captured within and by extensible, knowledge graph-powered data models, with all context on each data asset fully present and available, in machine-understandable form. With a data fabric, people and algorithms can make better decisions, while also reducing the likelihood and risk of data misuse or misinterpretation.

2. Answer Tough Questions: Data fabrics deliver answers via powerful query, search, and learning capabilities. Rather than a static entity based on moving or copying data, a data fabric platform provides a dynamic “queryable” data layer that gathers answers from across data silos. Previous data integration strategies relied on creating a new data model to support each new use case and then moving or copying data to fill out that data model. With a data fabric, data models are reusable, so when unanticipated questions arise, it’s easy for teams to adapt to meet the business’ needs.

3. Support Cross-Functional Data Management Projects: Data fabrics weave together existing Data Management systems, enriching all connected apps. They replace older systems that collected or cataloged an enterprise’s assets but failed to make the data usable. Previous solutions also failed in part due to their inability to handle hybrid, varied, and changing data but also due to organizational pushback. Data fabrics, however, are built for collaboration, leveraging and connecting existing assets, and driving a new breed of cross-functional Data Management projects.

Modernize Existing Investments

Most of us will recall how data lakes once held the promise of centralizing an enterprise’s data assets. But many data lakes fail to deliver on their hype precisely because they collocate data at the storage layer rather than connecting it at the compute layer. They leverage data based on its location rather than based on its business meaning. The whole premise behind a data fabric is that physical collocation of data does not by itself accomplish data connection or provide meaning or context. Older generations of storage-based integration systems such as the data warehouse are, in fact, even less capable than data lakes, since they only easily manage structured data to begin with, leaving the semi-structured and unstructured data silos completely unaddressed and disconnected. Companies quickly turned to data catalogs to try to address the bewildering diversity of their data landscapes only to learn that cataloging alone doesn’t lead to a connected enterprise.

While these technologies promised to end data silos, the truth is they are inevitable and exist for very good reasons. They allow for local control and governance when it is important to a particular part of the business, as some data must be stored apart from other data to comply with legal regulation or simply for legacy business reasons. Conventional data integration focused on eliminating silos through mastering, migration, consolidation, or governance. But data fabrics offer a practical alternative. Rather than working against data silos, a data fabric leverages them without requiring further copies of data. Instead of replacing legacy technologies, a data fabric works alongside existing investments and improves their utility. This is because a data fabric is an architectural design that operates at the compute layer and focuses on connecting data wherever it resides and, thus, actually improves existing physically consolidated data storage assets like data lakes, data catalogs, warehouses, MDM, and others.

Knowledge Graphs: The Missing Stitch to a Successful Data Fabric

Knowledge graphs are able to represent the full diversity and complexity of enterprise data because they serve as a universal format for meaning, regardless of data’s source structure, location, or format. A knowledge graph replaces the current laborious process for integrating enterprise data, which typically involves extraction, translation, modeling, mapping, and then moving data between various applications. The custom code required for modeling and mapping quickly becomes unwieldy at large scale, slowing the pace of innovation and insight.

Knowledge graphs are an integral part of an effective data fabric, as they create a reusable network of knowledge and easily represent data of various structures and support multiple schemas. Creating a queryable, reusable semantic understanding of enterprise and third-party data, knowledge graphs serve as the core of the data fabric: enriching and accelerating existing investments and providing critical access to business insight.

Just like an ordinary fabric that conforms to whatever it envelops, an enterprise data fabric lays over existing data assets and connects to them via individual threads and weaves these sources together into a unified layer. By doing so, data fabrics actually compound the business value of existing investments.