Disrupting the data landscape again with linked open data
Industrywide modeling
Other than the immediacy at which it enables data to be exchanged and comprehended, the capital value proposition of linked data is the singular continuity it provides for long-term investment in data-driven processes. The uniformity of its standards all but makes obsolete—although it can certainly work in conjunction with—legacy systems and the habitual need to overhaul them with more modern infrastructure that alienates time-honored data. In addition to rapidly becoming the de facto means of publishing data for public consumption, linked data is garnering even more traction within organizations in the private sector for those long-term benefits. A primary driver for that movement is the increasing prevalence of vertical industry standards—a definite means of not only accessing and analyzing data apropos to a specific vertical, but also of exchanging data between organizations (and perhaps even customers) within it as well.
Industry Foundation Classes (IFC) is one of the standardized ontologies within the construction field that is responsible for the exchange of data between countries in the V-Con project. Additional industry-specific ontologies include the Clinical Data Interchange Standards Consortium (CDISC) standards for clinical trials and the Financial Industry Business Ontology (FIBO) in financial services. “We’re definitely seeing a growing interest in industry standards and in FIBO in particular,” Polikoff says. “In financial services, many companies have one silo for a business glossary, another for security purposes and others for everything else. There is an emerging need to connect them and one of the drivers is regulatory compliance.”
Linked enterprise data
The realities of regulatory compliance, coupled with the increasingly strident penalties for non-compliance, serve as the primary impetus for the implementation of linked enterprise data within the private sector. Quickly realized gains in that regard include improved provenance and data lineage, as well as holistic management of information assets and their use according to industrywide regulations. Furthermore, adherence to governance practices is readily discernible for more effective data stewardship that provides additional oversight for compliance issues. “A lot of large banks are now using a linked data approach for data governance and data management, mostly because they are accountable for so many regulatory requirements,” Polikoff says.
The eminent expressivity of linked enterprise data delivers even greater yield to organizations with its innate understanding of the relationships between data, especially when that data represents the totality of an organization’s information assets. The standardized modeling of all enterprise data on a semantic RDF graph delivers peerless insight into the most minute relationships between data elements, enabling organizations to glean a pivotal contextualization of data assets that might otherwise seem unrelated. The result is a greater understanding of one’s data realized through a revamped data discovery process delivering analytic profundity through the use of all enterprise data, which positively impacts the ROI.
Machine-readable
Another fundamental trait of linked data that provides considerable utility within both private and public sector applications is its intrinsically machine-readable nature. On the one hand, that quality directly translates to an ability to scale—at rapid velocities—on sets of big data that might otherwise prove too exorbitant to manage. “The means to scale is one of the foremost advantages of this approach because it accelerates processes that would otherwise take too many financial and temporal resources to do,” Polikoff says. The underlying semantics technology that makes linked data machine-readable is also a core component of machine intelligence and artificial intelligence, enabling linked data to seamlessly merge and incorporate such data into applications of choice to the enterprise.
Silo unification
Of all the capabilities for linked open data and linked enterprise data to impact the future of the data sphere, the most irrefutable may well be the penchant for permanently abolishing silos. Linking data is far from synonymous with granting all systems access to each and every coveted node; governance and even security protocols can be implemented according to semantic triples to limit who can view what in accordance with organizational dictates. Still, the greater utility comes from the means of data interoperability among diverse systems at a pace equitable to that of the modern business climate, whether in the public or private sector. The allocation of resources for maintaining, modeling, integrating, transforming and preparing data for individual legacy systems or specific data marts is much too costly when compared to linked data’s efficient alternative of seamlessly harmonizing data between them. The additional benefits of scale, speed, governance and regulatory compliance, longstanding sustainability and machine identifiers certainly make this methodology primed to conquer the copious amounts of unstructured big data with which the world at large is contending.
Perhaps the future of data management may not result in ubiquitous computing or pervasive computing in which all IT systems are linked, but ascending credence is attributed to the notion that it will certainly entail linked data.