What is I-CONDA?

Digital transformation is introducing a growing number of connected devices that generate and consume a large variety of services from the Internet. To reduce latency and data traffic, these services are being shifted from large data centers to peripheral servers in what is known as Multi-Access Edge Computing (MEC). This shift opens key use cases for future networks, such as healthcare, agriculture, logistics, etc. Whereas this ambition is in the spotlight of larger urban areas, the slow deployment of 5G in sparsely populated areas is hindering the ability to bring MEC applications to these locations, which negatively impacts the development of industry and social development.

A solution that mixes multiple Radio Access Networks (RANs), including Wi-Fi, would bring MEC capabilities to areas where high-speed connection (e.g., 5G) does not exist. The O-RAN architecture has set the first step to bring edge and cloud components to Wi-Fi networks, enabling access to MEC resources through this consolidated technology, and embracing Artificial Intelligence (AI) to serve management purposes, such as radio configuration. In scenarios with limited or fluctuating network resources, adapting MEC configuration to the network status becomes key. This is especially true for delay sensitive services such as video, which is present in many MEC applications, including e-health, droning and remote assistance. Consequently, predictive quality of service algorithms and adaptive streaming mechanisms are key enablers to improve the quality of experience and the use of network resources. Thus, it becomes important for these locations to have access to such edge services, not only through 5G, but also via hybrid solutions based on more reachable technologies, such as Wi-Fi.

These environments bring added challenges as they involve distributed computational and network resources, which are complex to handle by human operators. Recent works argue the need of incorporating AI to optimally orchestrate such systems, anticipating changes and issues, and providing greater user satisfaction and performance, especially for delay-sensitive applications. Conversely, this heterogeneity and distribution of computing and network resources is expected to be part of Beyond 5G networks, which would have a greater level of resource geo-distribution, and advanced orchestration and management subsystems enabled by AI. Therefore, exploring advanced monitoring systems that enable zero-touch

management becomes of key importance to enable self-optimized, self-monitored and self-healing network operations.