IEEE Future Directions
Related Articles
Of the four approaches to Digital Platform for the Digital Transformation of Manufacturing, Industry 4.0, the one based on data is the one convincing me most.
The end-to-end approach is nice but probably over ambitious and it does not take into account legacy. The Communication based approach to me is just a component that would not help unless there is a parallel shift in a data based approach. The cloud, similarly to the communications, is sort of given, is the screwdriver that you need but what is really important is understanding how to use the screwdriver, where are the screws and which one need to be turned (the screws, obviously are the data).
At the same time, I noticed the huge, overwhelming variety of data, of ontologies and of structure of existing data. Each of them has a reason why and investment have been put on creating those structures and relative applications. Although this might be considered a “sunk” investment it creates a huge inertia making any change difficult.
This consideration applies beyond Digital Platform for Industry 4.0, you can see that it more generally applies to most verticals, from health care to smart cities, from logistics to constructions… The reason for discussing Digital Cities and Industry 4.0 was to show how different verticals may have different requirements and constraints (i.e. Smart Cities are driven by public investment, Industry 4.0 by private one…) but also to show the similarities and the unifying point is surely “data”. And, as I said, data are heterogeneous, they are tied to legacy systems, they have issues related to privacy, ownership, reliability….
How can we move forward keeping data at the core of evolution, yet taking into account their variety and the issues related to data?
It makes sense, when planning for the next step, to look further down the lane to understand the direction both in terms of where we would like to go as well as in terms of what directions are enabled by technology evolution (of course taking into account societal and economic implications). This is what FDC is really doing, taking a longer view span, identifying the desirable horizons and together with IEEE OUs taking the first steps in the desirable direction.
As an example: what is going to happen, or become possible, after Industry 4.0?
SAP have their own ideas and they are surely worth considering. According to SAP, see the graphic, by leveraging on the results of Industry 4.0, that make production more efficient from the market viewpoint by including the users in the value chain (getting direct feedback from product/service use through IoT and a flattened value chain), it will be possible to evolve the products/services much more rapidly (and I would guess charging the customer for the evolution). This can be done by splitting the atoms from the bits in the product. The product becomes a combination of hardware with a long life span and software having a much shorter (measured in days, sometimes) life span. The hardware part may be sold using conventional business models, the software part mat be sold as a service (paying for each subsequent release, paying a subscription fee, paying an on-demand feature evolution…). This is what they say will foster an “Incremental Innovation” with a direct customer involvement (empathy).
The subsequent phase fully blurs the boundary between customers and producers leading to an ecosystem driven innovation (by the way this is being hinted by the EIT Digital 2020-2022 Strategic Innovation Agenda) and this will lead to a significant disruption in the market place since the ecosystem takes the lead, at the expenses of today’s incumbent players (remember the loss of value implied in the shift from atoms to bits!). That is why SAP name this phase as Disruptive Innovation.
How can we take into consideration the diversity in data we have today and move towards an ecosystem where every player can potentially interact, make use, leverage on data?
I think the answer lies in the word semantics. By using semantics you can skip the issues related to syntax (which does not mean you can get rid of syntax!) and by operating on semantics different players can interact with one another. Of course you need to extract semantics from syntax and this requires in the easiest case some ontology, but in most cases it requires intelligence to correlate data into a context (which means with other data, with a time line, with a purpose).
For this we have artificial intelligence in its various forms, deep learning, machine learning being the crucial ones in this area.
What do you actually get when you start operating on data based on semantics (let me re-iterate: a semantics depending on the data-context relation)? You get Digital Twins!
Indeed, Digital Twins are a semantic representation describing and object, its behaviour, its history and its possible set of relation with the context in which it can operate.
Companies like General Electric have started to create and use Digital Twins as a springboard for their Digital Transformation and the Digital Platform they have created are based on Digital Twin interactions.
This, to me, is going to be the future of Digital Platform, the strong enabler for the Digital Transformation and .. the result of a successful Digital Transformation.
