Device Centric Distributed Compute, Orchestration and Networking

Abstract

The 6G wireless technology is expected to support extreme communication requirements in terms of throughput, latency, and reliability, as well as seamless interconnection of every object, processes, and people.The connectivity will be expanded to include wearables, bio-implants, intelligent machines, and vehicle components.To effectively represent the real world in a digital world for augmenting intelligence and making smart decisions, the vast computational needs must be widely distributed among devices, access points, edge, and central clouds.The ambitious 6G vision and envisaged use cases translate into rigorous performance requirements, which may lead to demands in data rates above 100 Gbps, fraction of millisecond (ms) latencies, and reliability in the order of 10 -7 to 10 -9 , which are beyond what is currently supported by 5G.Such extreme performance requirements can only be met by deploying necessary network infrastructure components, along with the intelligence and decision-making capabilities, to the very edge or in the proximity of the user. 6G Subnetwork ArchitectureRecent 6G literature [1-4] provides a plethora of unique technologies for the support of the demanding 6G services, with "subnetworking" being one of the major subjects.The main objectives deploying subnetworks consist in offloading most demanding services from the classical macro networks, to support extreme performance requirements (e.g., sub-millisecond communication cycle, wired-like reliability), at any suitable location and at any time.A subnetwork is a network of several devices and one or more Access Point(s) (AP(s)), which is (are) connected to one or more overlay networks, such as a 4G, 5G, or 6G cellular system, or a Wireless Local Area Network (WLAN).Subnetworks are anticipated to be components of a larger 6G system, while the AP is a gateway for edge clouds acting partly as a mobile device/User Equipment (UE) from the perspective of a 6G system.If no overlay network is available, a subnetwork can operate autonomously.A typical subnetwork architecture has a hierarchical structure (i.e., star or tree topology), where an AP controls the operations of the connected devices.For applications with non-extreme latency requirements, some of the connected devices can eventually act as relays and forward traffic to other devices.It is depicted in Fig. 1, with a representation of a variety of non-critical, medium-critical, and highly-critical latency applications.Subnetworks can also support replacement of wires among components with wireless connectivity.A typical scope of the subnetworking is introduced as "in-X" subnetworks that support extreme communication demands as autonomous, highly specialized cells with limited coverage to be deployed in environments where high-performance requirements are needed, such as in-factory, in-production modules, in-vehicles, in-avionics, in-home, in-robot, or in-human body area networks, for critical functions, like heartbeat control [1][2][3][4].