IEEE Future Directions
Related Articles
Exploitation of knowledge in the cyberspace
A CDT at stage IV becomes a knowledge entity that can be sold as a whole or by granting access to a subset of its knowledge space, through APIs.
A CDT can advertise itself, its “executable” knowledge in various ways and there will likely be knowledge platform for making this possible and managing the exploitation. As previously mentioned a CDT can share knowledge through interaction or by becoming a component in a higher level CDT (aggregating several CDT under one envelope).
As an example, a consulting firm can have a company CDT that aggregates its consultants CDTs and offer a single point of access to knowledge. Think about a consulting firm in the area of construction. Rather than offering single APIs to provide knowledge on building structural design, a different one for legal support, another for identifying suppliers, another for access to cadastral data (each of these is a complex area for which the firm has pool of specialists) the company can provide a single set of APIs to the firm CDT and it will be this CDT task to connect with the appropriate CDT.
The idea is to create clusters of knowledge to interface at higher level of abstraction, letting the CDTs orchestrate (with a platform support) the specific interactions. This will work both ways making it possible to return knowledge at the level of abstraction needed by the client. This increase the economic value of knowledge.
Notice that the networking of CDTs within the cyberspace can be flanked by applications that work on the knowledge exchanged and that may produce further knowledge.
A typical application case is the connection between the layers of
- Factual knowledge
- Diagnostic analytics
- Predictive analytics
- Prescriptive analytics
previously discussed.
DT can capture the factual data sent from their physical twin. These data can be transformed into knowledge –diagnostic analytics- (either inside the DT hence a CDT, or by an external AI application that feeds a CDT).
The next step is to have the CDT interacting with other CDT to correlate knowledge and extract/create further knowledge, some of it in a hypothetical for (what if), – predictive analytics – and finally taking decision on the available set of knowledge and release the one that matter –prescriptive analytics.
All of this happens in the cyberspace and different architectures can be used to support it. How much to embed into a CDT and how much to leave to an application is an architectural decision that has ownership aspects implied.
We are still in the first stages of research in this area to define and propose an architecture for CDT, although this aspect is addressed in the KaaS initiative.
