Data service factories

At the beginning of the composable edge idea train, there was a factory-like model which was intended to serve as the basis for a transformation engine.  Based on some earlier work on digital transformation outcomes, one might imagine the lower right of the matrix being research and development, the lower left being a production function for digital services, the upper left being a supply chain and the upper right a customer experience space.

The idea for this type of representation of a producer-consumer cycle isn’t new anymore, and it probably has many origins in both the natural and social sciences.  Which is enough to say that it’s probably in the public domain as a cycle of change model.

In working through materials on cloud-native computing, it has occurred to us that cloud-native devops is different in nature and kind from previous incarnations of system development.  We would either have a model (typically from some kind of mathematics) or experience building a similar thing to what was on offer to be built next.  Timelines ran to years; team sizes were in the scores or hundreds.

None of that is true, anymore.  Individuals can outperform a large disinterested group without much trouble, for one.  Timelines for delivery of working examples are in the days or weeks range rather than the months or quarters.  Models can be built from data on-hand, even if that data is less than the amount nominally statistically valid for assumptions basing.  So much of this is good we have to wonder what’s not so good.

And we are not going there, hopefully ever.  If we can do more with less, and we can leverage the more with less again and again, we get a kind of fractal growth model that is chaos at work on behalf of some greater good than on behalf of finding the darkness in stings already built.

A chaos engineering for enlightenable change, along with a perpetual motion machine to continue its gifts on and on into betterment realms.  There are opportunities to extend available knowledge in all matters of direction, if only we had some.  It could be that the remnants of pandemic have made thinking about the future less attractive than it once was, but maybe a step-wise approach can help.

Let’s say there are five axes of importance, and we want to track them in the form of equations of motion.  For talking purposes, the axes are the domains of a digital transformation theory.  These are: a digital business, its digital policies, its digitalized processes, a digital workforce and a digital information infrastructure.  Here, we are going for all digital, all the time, including getting to digital first as well as digital next.

That is, we are looking for an engine which can not only self-heal and self-correct if it gets off-course, but one that can also self-migrate to better positions relative to productivity, sustainability, and lifecycle management.

There are already ways to accomplish this and we can put some of those ways together to get a steerable change control system.  Self-steerable, that is.

Starting with a Data Service model, the matrix below provides a foundation for formulating equations of motion.

The general pattern of Management through Models provides a mechanism for organizing a data services delivery system.  Each level of attainment provides both a substantively different perspective on the data under management but also incremental meta-data that can serve as the co-ordinates of change.  Each Modality of the data can be vectorized by the NIST Big Data framework of volume, velocity, variation, variability, and value.

In a second example, the Value Chain models are data pipelines which reach from the research and development teams to a customer experience.  Building the value chains is similar to the process which links electricity generation capability to residential and commercial users.  The model for that process is:

Smart Grid: Definition, Goals, Objectives, NIST Conceptual Model | Electrical A2Z; accessed 15 august 2023, rights reserved to owners of the material

In this representation, the lower right of our first projection is now on the lower left of the Smart Grid representation.  Alignment aside, we could structure the movement from the lower right to the lower left in the matrix above to be ‘upstream’, and the motion from the lower left to the upper left as ‘further upstream’.

Which makes it possible for us to use a data mining model as the representation of the upstream reference architecture.

Based on a version of the CRISP-DM methodology, the circular use of data derived from the use of open standards to manage a digital transformation provides a scalable architecture for the purposes here.

This provides the cyclic behavioral model that can be used to support the construction of a customer catalog through the adoption of the IT4IT version 3.0 digital system devops model.  This is presented in the graphic below:

Here, we have value chains at the level of Smart Grid Domains (from the SEI Smart Grid project) applied to the creation of customer outcomes.

As to managing the entire process of processes, we identified that methodology in a previous note.  Model Reference Adaptive control of the entire process of generating digital transformation capability can be represented as:

Specifically, in the instance of a model for digital transformation via the use of the IT4IT model can be represented as:

The digital transformation process, then, becomes a series of representations, linked together as the supporting documentation for how the digital equations of motion determine the usefulness of reference architectures, best practices, use cases, and the case management of an organization’s digital transformation program.