Design Considerations and Guidelines
The selection of the right use case is driven by a detailed assessment of potential collaborative business models which are shared among the targeted organizations for consideration and consent of participation . Betti ( 2020 ) defined a five-part framework to help organizations :
1 . Understand the business challenges . 2 . Develop applications to help overcome the business challenges . 3 . Determine the viable applications . 4 . Identify suitable collaborators for each application . 5 . Define the data-sharing relationship .
For FedL , collaborations can be one of two models , sharing or allocation ( Man , 2019 ). In the sharing model , collaborators can work together on with a pre-agreed split to capture value and realize greater scale or network effects that the pooling of data can provide , that is , the value of the data and information gained from it will increase when it is aggregated and shared . In the allocation model , value is created by optimally allocating risks between collaborators thereby allocating greater profit to the partner managing more risk ( or more contributing data ) in a FedL alliance . For example , as ( Yang , 2019 ) has identified , FedL could make equitable rules for profits allocation with the help of consensus mechanism from Blockchain technologies .
At present , the primary approach to FedL alliances is via data sharing platforms provided by independent 3rd parties , who are not participants in the alliance . Such platforms may be attractive to organizations who cannot build their own platforms for data sharing or who want to outsource the challenges of managing data transactions , matching , licensing , and transfer of data ( Richter , 2019 ). These platforms also have the benefit of solving the problem of trust , not only with having the platform being provided by a neutral 3rd party , but also through features such as certification , secure access control and digital watermarking ( Richter , 2019 ). An example of this is the FedAI Ecosystem developed by WeBank , which offers an IoT driven data sharing platform , with applications in the areas of vehicle insurance , financial lending , and anti-money laundering .
Another scenario that utilizes the FedAI ecosystem is to create and online visual object detection platform to support collaborative object detection and minimize excessive data transfer to central cloud servers from cameras deployed in the field . Another example is IBM ’ s Maximo visual inspection that has presented such a federated system for vision-based defect detection in manufacturing products ( Bharti , 2021 ). Such Industrial IoT use cases can prove to be crucial in organization ’ s growth in a competitive environment where if the defected products are delivered to customers this could not only hamper the organization ’ s reputation and future business prospects but also result in significant losses in terms of raw material wastage and cost of recycling .
IIC Journal of Innovation 25