Sunday, July 13, 2008

Semantic IT Practices

As noted in a previous post, the current set of methodologies employed in the day to day IT operations of a typical enterprise is poised for perhaps its most significant paradigm shift in several decades. This evolutionary shift is not the introduction of Semantic technology or standards per se, but rather the complete re-visioning of how IT works in the context of Semantic Interoperability. Semantic IT provides us with two crucial capabilities that we simply never had before:

1 – The ability to abstract the governance and maintenance of the systems under our control in an automated fashion.

2 – The ability to link all the various disparate IT-related processes through that same abstracted Semantic governance layer.

What’s really missing at this stage are the semantically enabled methodologies and/or practices that will allow us to exploit these capabilities effectively.

So what exactly is a semantically enabled practice or methodology? The difference that Semantic Enablement brings to enterprise operations is the deliberate focus on operational fusion or merger of processes in an effort to achieve a holistic management paradigm. This means that there are no longer any “stove-piped” approaches or technologies, all efforts must have the capability to inter-relate with all other capabilities and share the same foundation.

This holistic approach will require several new or at least modified IT practices with updated methodologies designed to take advantage of both the philosophical and technical shift in thinking. It is likely that this should lead to a certain level of consolidation in the number or diversity of existing IT practices as the central tenet involved is essentially holistic awareness and interoperability. Once it is realized that all aspects of IT architecture and operation are in fact related, the need for diversification and specialization will be reduced. There should not for example, need to be separate disciplines for Master Data Management and Semantic Integration, as MDM represents a specific application of Semantic Integration across multiple data sources. In fact, many or most aspects of data architecture or integration could at some point be considered within the umbrella of Semantic Integration.

One of the real benefits of viewing IT methodologies or practices this way is the deliberate attempt to keep them technology agnostic. These practices are not driven by vendor solutions. Previous practice approaches based upon individual or categories of vendor solutions have often proved short-sided and counter-productive, leaving the enterprises which adopted them vulnerable in many ways. The application of operational IT capability (i.e. business services) doesn’t require specialization of support infrastructures, instead it should have a flexible common framework that can be redefined as needed. Our goal is not to be able to predict all possible permutations of our solutions and build that into some rigid roadmap – the true goal is to give control of that map to those who need to adapt their solution in their own context, according to their needs.

That foundation then is an abstracted governance or management layer which mediates between all others aspects of IT architecture. It is accessible to and modifiable by stakeholders, which allows for much more economical and efficient governance processes. One of the most exciting parts of this is the opportunity we now have to merge design, architecture and governance using semantic modeling. This will likely manifest itself in a blending of ontology management and EA framework or even agile modeling techniques.

I’ve taken an initial stab at defining a set of Semantic IT practices, these have all been developed to replace existing counterparts and designed to be interoperable with one another. These practices encompass most if not all current IT processes and include:

Semantic Integration (most generic in nature, next generation solution for enterprise systems integration).

  • Service Oriented Integration (service & application logic).
  • Program Lifecycle Management (program, product, project, portfolio, process synergy).
  • Dynamic Learning Orchestration (organizational and personal learning and knowledge management).

In my next blog entry, I will describe Program Lifecycle Management, the new PLM, in more detail.

Semantic Integration & IT

In many ways the practice of information technology has changed little over the past 30 years or so. It may not seem so on first appearance - but the premises upon which our current technologies are still operating are largely based on philosophical constructs that date back 30 years or more.

Those constructs include:

  • The Relational Database.
  • The Data Warehouse.
  • System Management (system as self-contained entity).
  • "Static" Business Rules Management (more or less permanent view of the nature of rules within the enterprise) - exemplified by maturity models.
  • The notion of IT service provision as separate and distinct from the elements of the business which they serve.

Of course, there are exceptions to this but in many ways these basic foundational elements still represent the core of most IT activity within the typical enterprise. Each of these areas and others has arisen to meet certain needs; there has been an ever-increasing level of specialization within IT that has more or less extended the basic model without really changing it. This has made some areas of IT management more effective but has hindered the ability to truly unify IT capability. Some people refer to these phenomena as 'sub-optimization' or 'compartmentalization.' In the government arena it is sometimes referred to as 'stovepiping.'

The dilemma is the same though, how does one make progress and get work done while solving near-term specific problems and simultaneously leverage the full potential of all enterprise assets? This can be done and can occur in the near-term using existing technologies - the problem we face now though is one of perception and philosophy. IT is used to working in its current sub-optimized mode, its relative detachment from the primary functions of the organizations it serves tends to provide a disincentive to view the big picture requirements they always receive yet never truly fulfill.

The paradigm shift that needs to occur and that will occur within the next five years is this - The realization that the enterprise is a dynamic entity. In other words, it is not something that can ever be fully coded or captured in advance. A 'pre-determined' enterprise family of systems is obsolete before it ever deploys. The other element of the realization is that the key to all integration within the enterprise is already embedded within it. The enterprise as a holistic capability will be achieved once we understand that regardless of the specializations involved, all aspects of the mission are related and that the unifying code embedded throughout it can be managed, mediated and coordinated as a single semantic exercise.

I truly believe that this represents more of a philosophical change than a technical one. It requires new methodologies and IT practices that view problems from this updated perspective. Eventually it will lead to modifications in many commercial software products and networking technologies, but those changes are not necessary to begin experiencing the benefits of the Semantic Enterprise.

Copyright 2008, Semantech Inc.