What is Reference Architecture?

Reference Architecture is a systems & solutions architecture which depend on Reference Data within a generic Reference Model of business information. The focus of GlencoIS is Reference Architecture based interoperability and the implied architectural requirements for full life-cycle solutions, independent of any given tools & technologies. Quite simply, whatever business and technologies, and however these evolve, there are dependencies on how shared information is referenced and linked.

ie Whatever your physical systems or solutions architecture it should always honour interfaces to both tools and technologies that maintain independent referencing to your required business information standards libraries:

[Note: The technology landscape and the fashionable business buzzwords evolve ever faster – whether Integration or Interoperability, A/LCIM or BIM, iRING or CFIHOS, RDL or MDM, Cloud or Everything-as-a-Service, Digital Twin & Digital Transformation or Blockchain, AI or AR, you name it – life-cycle solutions continue to depend on key Reference Architectures.]

Why should I care about Reference Architecture?

You already recognize the great business value in achieving flexible interoperability between business-units, supply chain partners and systems, over long multi-project, multi-asset, multi-business, greenfield and brownfield life-cycles, through:

  • direct cost-and-time-savings in reducing effort in transferring, mapping and in simply creating, sharing, finding & accessing information necessary to do your business.
  • risk-and-cost-reductions in the quality and ambiguity of information and management meta-data which otherwise lead to sub-optimal business operations, failure to satisfy regulators or, in the worst case, loss of health, safety & environmental integrity and security.
  • freedom-to-take-advantage of new opportunities in collaborative business processes, flexible business partnering, and different subcontracting arrangements across your evolving, remotely distributed, business operations and supply-chains. In a context of collaboration and trust, that same evolutionary flexibility enables agile working in creation and deployment of new systems and solutions.

And you already recognise that achieving these benefits has a strong dependency on standardisation across systems, across information and across technologies. System standardisation has intuitively obvious benefits, but standardising the wrong things can create a millstone – costly constraint on future business – so it is important to understand the right things to standardise at the right level. Standardisation is a system architecture issue.

Different business domains already address different parts of this problem with varying degrees of success.

What is clear is that drivers for information & communications technology standardization are much broader than your own business. And it is also clear is that software solutions & services have quite different supply-chain development and support life-cycles compared to your core business.

That is …

  • Business Processes & Applications
  • Data & Information
  • Implementation, Communications & Security Technologies

… must be standardized independently wherever possible, if the benefits are to be realized. This is the key architectural requirement – whatever the technologies.

ISO15926 is a standard built on this architecture, focussing on standardization of information (models and semantics) independent of particular business applications and usage, and neutral with respect to implementation technologies. It is far from being the only industrial standard that recognizes such a focus on the standardization of information in data dictionaries or reference catalogues. Counter-intuitively, details of the information ontologies and libraries standardised are much less significant than the overall architecture implied. How reference data is used is much more important than exactly which reference data is standardised.

So, a key feature of ISO15926 is that shared information definitions are handled as Reference Data; reference data that is itself accessed and managed independently of your business information and its implementations, and indeed independently of the many possible standards and catalogues that may define that information.

The other key feature is that all information, including reference and master data, is modelled according to a so-called “high-quality” Generic Entity Model, one that supports a long business life-cycle beyond currently well understood and predictable business process arrangements.

These information architecture reasons are much more important than the choices to use specific ISO15926 or BIM resources, since these, like all the other elements, continue to mature and evolve. The following case study is 15926 based, from JORD 2012.

ISO15926 Reference Data supports
interoperability & sharing of information definitions
between different information standards
as well as between different businesses & technologies.

Because the business scope of such wide interoperability is very large, it is impractical to expect the critical operations of any major business, let alone whole industries, to rely on a single reference data service provider for life. A key feature of the ISO15926 RDS architecture is that it also supports multiple federated reference data systems and operations because, being entirely technology neutral it can also exploit the most standard “semantic web” implementation technologies without imposing any dependency. The reference data available to any one business is the sum total of a whole web of linked resources. And, because the way those linked information resources are represented is itself standardized, they benefit from sharing a common scheme for their quality and compliance validation.

What does it take to exploit Reference Data practically?

The business benefits and attraction of succeeding in adopting the ISO15926 Reference Data-based Architecture for interoperability ought to be intuitively obvious given the description above. But the huge scale and complexity would lead anyone to be wary of seeing it as a single viable implementation project or business. A metaphorical elephant. No-one should try to eat a whole one.

[2019 Update: ISO15926 Reference Data has been a series of collaborative projects since 1991 though is itself built on ISO “STEP” / EPISTLE and StepLib in initiatives from the late 1980’s. From 2011 to 2013 the Joint Operational Reference Data (JORD) project was a collaboration facilitated jointly by PCA (POSC Caesar Association) and FIATECH to establish a core set of services, methods and capabilities. This core enables the evolution, management and validation of compliant use of reference data not only by your business directly, but by a federated web of competitive providers of value-added content, systems and services to your business. These developments continue today in initiatives like CFIHOS & MRAIL (by USPI, PCA, IOGP etc.) open to members and non-members alike, anyone who recognises that their achievement of lifecycle interoperability depends on shared reference data.]