RESILINK NEWSLETTER #1

To implement the generalizing local resource sharing approach, RESILINK develops a distributed digital resource management platform for real-time exchange of information on territorial resources, supplies and demands; connecting smallholders to new supply, sharing opportunities and distribution channels. Although RESILINK shares many functionalities of selling platforms, its purposes are to integrate and test innovative approaches. For instance:

• RESILINK will deploy a resilient network ensuring high availability of services where a number of light-weight platforms can be installed on local servers hosted at regional or city-level or even at the community level by socio-economic organizations such as Chambers of Commerce, cooperatives, government agencies, start-ups/SMEs, etc. The platform can also be installed in a distributed way so that several platforms can operate simultaneously to manage resources according to geographical areas. Each of the distributed platforms can share the information on territorial resources.
• RESILINK will adopt a Blockchain approach to ensure robustness of the decentralized transaction system. Blockchain technologies have been investigated in a large variety of applications and RESILINK, with its private and decentralized architecture particularly investigates how private/permissioned, peer-to-peer and hybrid Blockchain frameworks provide a possible solution to problems with a single point of failure and bottleneck.
• RESILINK will implement a platform-of-platforms approach, capable of integrating existing or future platforms into comprehensive dashboards/portfolios. An open API will be developed and made available for third-parties to benefit from the software building blocks in order to quickly develop other resource sharing applications that can all interoperate, thus extending the ecosystem.
• While the integration of AI technologies may seem nowadays easier with the increased maturity of the research domain, the traditional data models in AI usually assume a simple data exchange model. In these traditional models, one entity would build and transfer sets of data to another entity which is in charge of cleaning and fusing the data. The AI data processing chain could include additional entities, each providing at its level its own AI approach to build new models for other users. When implemented on top of a highly-distributed architecture, with the constraint on secured data management, this traditional way of running AI has to face data fragmentation and data isolation issues. Recently, decentralized AI approaches have been proposed to tackle these issues and RESILINK will investigate federation-based AI or gossip- based AI approaches.