By Dr. Jans Aasman
The demands for contemporary data governance are greater than ever. Adding to the pressures of mounting regulatory mandates (and penalties) are e-discovery concerns and the increasing decentralization of the data landscape within and outside the enterprise.
The resulting complexity places a premium on understanding where data are, who has access to them, how they’re moving, and how that relates to multiple regulatory accords at the precise time it’s most difficult to do so. Implementing the protocols, practices, and people for effective governance hinges on this understanding, which becomes even less clear when outlining these responsibilities once IT systems are in production (which is typically the case).
Attempts to redress these complications commonly involve hiring consultants or investing in governance products that provide point solutions with fleeting efficacy. However, organizations are starting to realize that by dedicating these same monetary and temporal resources to an enterprise knowledge graph, they can perform many of the same functions of these point solutions—and more—by themselves for unparalleled visibility into their governance needs.
The difference is the latter approach is far more sustainable, comprehensive, and flexible. By connecting all data assets with a linked data approach, organizations can perform even ad-hoc impact analyses to determine just how data, their sources, and users correlate to workflows to first determine, then strengthen, governance measures.
An internal enterprise knowledge graph details how IT systems and their data are interconnected. It encompasses the full range of IT systems involved in business processes and links them together, regardless of how distinct they are.
The result is a comprehensive graph of internal data resources spanning business domains, use cases, and applications. Such a knowledge graph not only illustrates the relationships between individual data nodes, but also the different systems with which they interact. This graphical representation is ideal for holistic impact analysis of the movement of data, creating the starting point for identifying what governance measures are required.
Moreover, it does so with a remarkable degree of specificity. Users can look for data related to business terms and quickly see all the databases in which they’re stored. The conflation of this visibility and the detailed provenance capabilities of self-describing smart data identify just where data go on their journey throughout the enterprise, who accesses or alters them, and where they ultimately end up. It identifies everything governance councils need to understand data’s location and movement, forming the basis of pivotal planning requirements.
Enterprise knowledge graphs are equally valuable for both implementing and enforcing policy. They show data stewards exactly how data has progressed through an organization at all of its different levels. Critical to this functionality is the visibility into how IT processes relate to business ones. Thus, organizations can proactively see where data goes to implement policies around who can access or modify them.
Retrospectively, stewards and IT teams can rely on fine-grained data lineage capabilities (supported by detailed metadata) to determine any anomalies or policy breaches. This insight is critical to structuring specific use cases in accordance to governance rules. For instance, when ensuring that sensitive data complies with regulatory requirements, organizations can use this knowledge to certify that each of the zones traversed by this data is impacted by those protocols.
These same provenance capabilities are an effectual means of demonstrating regulatory compliance too. Therefore, internal knowledge graphs clarify governance from initial planning to operations, whether that involves satisfying regulations or simply maintaining data as an organized asset.
The First Step
With the proper visualization mechanisms, enterprise knowledge graphs are able to first illustrate where data are in an organization’s IT systems, then pinpoint which governance mechanisms are needed for them via their linked data approach. Linked data, or linked enterprise data for internal knowledge graphs, enables IT systems of any types to be joined for the seamless transfer of data between them.
The underlying smart data techniques of enterprise-wide taxonomies, naturally evolving ontologies and semantic graphs deliver the framework for linking these resources in an all-inclusive graph. Organizations can create knowledge graphs by replicating their data into a semantic graph format and connecting the different repositories or IT systems; the more of these they include, the more visibility into their governance needs they receive.
Data Governance Blueprint
Linking data in an enterprise knowledge graph facilitates the sort of impact analysis needed upon which to base data governance. In this use case, the knowledge gleaned is potential regulatory compliance vulnerabilities, security issues, access needs, and the accordant governance responsibilities to maximize data’s enduring value across the organization. The chief advantage of this approach is its innate flexibility. Whenever organizations add sources or reconfigure business strategies or processes, they can always rely on this method to deliver the same visibility into any data governance concern.
Jans Aasman is CEO of Franz.com, an early innovator in artificial intelligence, a Ph.D. psychologist and an expert in Cognitive Science. As both a scientist and CEO, Dr Aasman continues to break ground in the areas of artificial intelligence and semantic databases as he works hand-in-hand with organizations such as Siemens, Merck, Pfizer, Wells Fargo, BAE Systems, as well as US and foreign governments.