6G Architecture to Connect the World

We missed Nokia researchers publishing a paper in IEEE Access, '6G Architecture to Connect the Worlds'. The paper explains many concepts nicely and is worth reading for any researchers involved. The abstract as follows:

We explore several novel architecture concepts for the 6G era driven by a decomposition of the architecture into platform, functions, orchestration and specialization aspects. With 6G, we associate an open, scalable, elastic, and platform agnostic het-cloud, with converged applications and services decomposed into micro-services and serverless functions, specialized architecture for extreme attributes, as well as open service orchestration architecture. Key attributes and characteristics of the associated architectural scenarios are described. At the air-interface level, 6G is expected to encompass use of sub-Terahertz spectrum and new spectrum sharing technologies, air-interface design optimized by AI/ML techniques, integration of radio sensing with communication, and meeting extreme requirements on latency, reliability and synchronization. Fully realizing the benefits of these advances in radio technology will also call for innovations in 6G network architecture as described.

We liked the 6G het-cloud concept shown in the picture above. The following is from the paper:

Cloud transformation towards 6G will have many and heterogeneous aspects. Hence, we have coined the term ``het-cloud''. In the following some of these aspects relevant for 6G architectural transformation will be described. Diversification of cloud-based service delivery platforms into separate private, public, on-premise and edge clouds call for closer co-ordination of distributed computing and communication resources through federated network control and orchestration.

The het-cloud environment is a heterogenous cloud environment with multiple stakeholders to run applications at different sites such as on-prem, far edge, edge and core with a variety of different hardware and software stacks. The clouds can be private, public or hybrid. There are two main benefits of the het-cloud approach. The first one is the ease with which new services can be created, placed, subsequently scaled and moved between the clouds and the efficiency with which they can be executed. The second one is the knowledge of the cloud capabilities to optimize service performance. Such a het-cloud approach is foundational in terms of flexibility and simplicity and well in line with 6G architectural expectations. Also, such a concept will allow highest level of trustworthiness by implementing trusted execution environments (TEE). TEE will guarantee the integrity and confidentiality of both code and states; remote attestation will provide proof of trustworthiness to third-party stakeholders.

The unit of execution may vary from a stateless function to a micro-service and to a full service in a container or a virtual machine in the het-cloud environment. The het-cloud will encompass cloud software platform for serverless functions in the `FaaS' or Functions-as-a-Service layer that sits on top of the Backend-as-a-Service (BaaS). The BaaS contains the network related operations and intelligence such as data collection and analysis, logging and monitoring and distributed data storage. Similar to how a UE can offload application execution to the edge clouds, service and network function execution in the het-cloud environment can be dynamically relocated between the connected clouds making up the het-cloud.

The implication of such dynamism is that the network functions and service must be implemented by movable and extensible code. Before relocating the complete service, parts thereof, or a constituent function, the offloading entity needs to discover the capabilities of the connected clouds. For this reason, the connected clouds need to announce their capabilities to their peers. In particular, the announced cloud capability should include exposure of available hardware accelerators such as GPUs and trusted computing platform modules of TEE with remote attestation support and APIs that are integrated into the cloud platform. Other examples of such capabilities include supported computing types (such as IaaS, bare metal, PaaS, SaaS, FaaS), high availability characteristics, processing capacity, latency and type of cloud platform. We thus envision a cloud capability discovery and function placement service as an integral part of the future het-cloud as shown in Figure 3 above.

The direct PDF link of the paper is here.

Related Posts: