The Problem
Full integration of all IoT devices is not practical. A unique point to point interface is needed for each pair of participants to communicate. The development and support of this unique point to point connection is limiting the level of integration that is possible as companies expand from an internal system to interacting with thousands of other participants.
Think of checking one’s bank account balances across several banks. With the current WWW technology, a user would be confronted with a bank specific portal as well as the unique proof of identity for each account they wished to check. Furthermore, if the user wished to transfer funds between banks there is additional complexity of getting those authorized funds from Bank A to Bank B. To automate this process, 6 one way point to point interfaces (or 3 bidirectional interfaces) would be needed between all the participants to communicate as shown below.

This configuration is manageable and can be further transformed by going to a hub and spoke configuration. This is the modern interface design which can reduces the number of interfaces that need to be developed and managed. The image shows 3 bi-directional interface which is equivalent to the 6 point to point interfaces as shown above. The Hub is then responsible for translating requests from the participants into the proper format for the intended/destination participant. It is clear in this 3 participant configuration, the addition of the Hub only increases the complexity between the active participants.

The Hub & Spoke design is more powerful in its ability to scale-up. It scales linearly where point to point scales close to the square of the number of connections N*(N-1). Below is a visual example showing 14 participants with the Point to Point (182 connection) and the Hub & Spoke (14 connections) topologies which provides a more reusable interface. That is, it can be leveraged for new participant connections.

This seems like a great topology to drive the 4IR. Given 1000 proposed ecosystems or unique participants, Point to Point would require 999,000 unique interfaces/connections while Hub & Spoke would simply require 1000 (bi-directional). Why isn’t this topology used across the internet?
The issue is the Hub would need to be trusted by all participants, worldwide. The complexity of the hub would grow similar to that of the point to point solutions if each connection had a unique data structure/API which needed translation. This Hub & Spoke approach works best when the types of messages are of a standard format, such as Financial Transactions, Material Movement, HL7 messages, etc.
Additionally, participants are usually manually connected to the Hub & Spoke systems (via IP address or UN/PW for example) due to needed authorization and authentication. Also, this structure requires the flow of all data & transactions through a central hub. This can cause data/transaction performance bottlenecks reducing the participants experience and responsiveness. All of these issues create a completely different set of challenges for a final technological solution.
What is needed to enter the 4th Industrial Revolution?
The hub & spoke model starts to give some hope of a solution but doesn’t completely fulfill a critical design requirement as stated by Klaus; it must be integrated and comprehensive. The design does seem to show good integration. Although, it does lack being a comprehensive solution as it relies upon the flow of similar data between all the participants to maintain a low complexity. Supporting different data streams would increase the hubs complexity as well as put the onus of developing and deploying the solution upon a central organization. Who would fund the central organization and how would it be governed?
This hub & spoke model is a good basic design, but its limitations need to be addressed; centralized design, API bottlenecks, static/manual topology, inability to authenticate and trust participants. Turning this statement around to what is needed; an efficient decentralized design with the ability to authenticate participants upon a dynamic topology.
To correct these limitations the centralized hub needs to be distributed. This can be accomplished by pulling a copy of the hub concept towards each participant, integrating all participants. These new hubs, labeled as UFT (Universal Framework of Things) in the following image, are responsible only for the data/API important to their interactions with other participants. This will also decentralize the data flow making the system more efficient.

To make the solution comprehensive, a few more things need to happen. The UFT(s) need to have an identity known to the other participants within this overall EcoSphere (group of participants & companies wishing to interact) which is achieved via a comprehensive trust model. The full comprehensive solution can be realized when the UFT is given the ability to automatically learn how to connect to other participants to dynamically change the topology of the EcoSphere. With this ability comes simplicity and opportunities for automation. Participants can publish their API (connection model) and other interested participants can learn that model with no programming or manual update. One last design consideration for the solution; security, privacy, and encryption is a design principal, not an afterthought. These things taken together deliver a superior topology that can be adopted by existing legacy systems and new solutions ushering in the 4IR.
More Information
The Digital EcoSphere described above is one of the solutions based upon NeurSciences’ technologies. The company has developed tools such as the Universal Framework of Things (UFT) and a MultiKey Knowledge Graph technology for a comprehensive trust model. This is all rolled into a Platform as a Service (PaaS) which enables automation of independent ecosystems.
For more information on this and other solutions with NeurSciences’ technology, please email info@neursciences.com.

