IDEA FOR AN AGENTIC AI DECENTRALISED RAN TELECOMS NETWORK

The idea is to design and test an innovative and more modern type of telecommunications network. Agentic AI and front-end wireless modules integrated with fixed or moveable modular data centres located at the core and the edge of the network would enable newer cutting edge applications to be developed while at the same time preventing network outages in the event of lack of power, a
cyber security attack or from unfavourable weather conditions. In this case data would be disaggregated across the data centres and base stations. These data centres would have energy storage capabilities and be connected to and communicate with a smart electricity grid powered by a
combination of renewable energy and fossil fuels. AI and machine learning would sense which data centres were having issues with power and/or water or which were having connectivity problems. It would be a totally flexible system where agentic AI RAN’s, smart-grids, water systems, data centres and end devices/sensor nodes would work in unison to process data and data would be routed to another centre if there was a problem with power, unfavourable weather, if it was being subjected to a cyber security attack or having a communications outage. Another area where this would be useful is to process data at a centre at off peak times where the load required would not be critical. Decentralised Edge/Modular data centres could also be used for soverign AI, keeping data within country or regional boundaries, thus adhering to GDPR regulations and Data Protection legislation. Ideally these data centres would be located on mobile base stations, in smaller municipal districts, rural areas, smart rural regions and along the coast. With AI technology still maturing, front-end modules from different hardware manufacturers including start-ups could be combined and/or interchanged as it evolves. Both Open Source and Domain Specitic language models (DLM's) that are application or regulation specific could be tested.