MACHINE LEARNING-BASED, NETWORKING AND COMPUTING INFRASTRUCTURE RESOURCE MANAGEMENT OF 5G AND BEYOND INTELLIGENT NETWORKS

Project Information

MARSAL

Grant agreement ID: 101017171

DOI

10.3030/101017171

Project closed

EC signature date 8 December 2020

Start date 1 January 2021

End date 30 June 2024

Funded under

• INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies - Information and Communication Technologies (ICT)

Total cost

€ 6 126 683,75

EU contribution

€ 6 126 683,75

Coordinated by

IQUADRAT INFORMATICA SL
Spain

Periodic Reporting for period 2 - MARSAL (MACHINE LEARNING-BASED, NETWORKING AND COMPUTING INFRASTRUCTURE RESOURCE MANAGEMENT OF 5G AND BEYOND INTELLIGENT NETWORKS)

Reporting period: 2022-07-01 to 2024-06-30

Summary of the context and overall objectives of the project

Targeted innovation activities need to take place to fully exploit key technological developments, towards a disaggregated infrastructure model, where technological infrastructure blocks can be transparently and flexibly replaced by others. Key advances are required both in the network design and network/service orchestration levels, which are able to support multiple distributed edge nodes, intelligent network management, and information security especially in multi-tenant environments. To this end, the MARSAL project aims at proposing a new paradigm of elastic virtual infrastructures that integrate in a transparent manner a variety of novel radio access, networking, management, and security technologies. This is achieved by developing innovations that are based on three pillars: the network design pillar, the virtual elastic infrastructure pillar, and the network security pillar
The MARSAL objectives are:
Design and demonstration of scalable, distributed cell-free massive-MIMO networks supporting massive AP deployments
Design and implementation of a cell-free vRAN for B5G, aligned with the O-RAN Alliance architecture
Architecture of a disaggregated, SDN control plane towards Fixed-Mobile Convergence
Deployment of an Elastic Edge Computing paradigm with Cloud-Native technologies
Policy-driven security, privacy and trust in multi-tenant infrastructures
Deliver of a Self-driven infrastructure with pervasive, ML-driven control
Implementation and proof-of-concept of the MARSAL solutions
Dissemination, standardization and exploitation of MARSAL

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The MARSAL partners have identified use cases and verticals, in the context of 3GPP, ORAN, 5G-PPP, and industrial initiatives, in order to determine the common aspects with the MARSAL Proof-of-Concept (PoC) scenarios. They have also initiated the procedure for the definition of the overall MARSAL architecture. Initial progress for the design of the cell-free RAN approach of the MARSAL project has been achieved. The cell-free networking concept has been aligned with O-RAN specifications at the PHY layer. Also, the definition of the system-level specifications and interfaces for the wireless connectivity for the Hybrid MIMO fronthaul for cell-free networking is achieved. MARSAL has achieved innovation on the application of AI-based algorithms for the Near Real-Time Radio Intelligent Controller, by considering the O-RAN specifications, as well as on the design of the MARSAL optical network configuration is also achieved. The design of the multi-objective optimization algorithms, targeting the joint allocation of wireless and wired resources has been successfully finalized. The design of privacy preserving pipelines to allow the sharing of confidential information for its use in data-driven Network Slicing applications, and the development of a blockchain-based Smart Contract platform, for decentralized Network Slicing transactions in multitenant infrastructures have also successfully implemented.Finally MARSAL has achieved a significant number of dissemination, communication and public engagement activities, through the project’s website and social media accounts, the publication in journals and international-conferences’ proceedings, brochures, newsletters, press releases and presence at industrial exhibitions.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Beyond the state-of the art:
Definition of the core architectural components and specifications of the novel MARSAL network architecture
Alignment of the cell-free networking concept with O-RAN
Analysis of a mmWave solution to provide inter-O-DU coordination
Development of novel radio resource scheduling algorithms
Design of a new RFIC that supports the 60 GHz standard
Initial design of a 2x2 60 GHz hybrid MIMO system
Design and of the MARSAL’s optical Midhaul transport network
Design of resource allocation algorithms in optical-wireless networks
Instantiation of a MEC platform based on ETSI MEC functionalities
Coordination of the Network Function Virtualization Orchestrator and the Multi-access Edge Application Orchestrator (MEAO)
Definition of the Service Function Chain problem for the MARSAL architecture
Design of MARSAL’s Decision Engine and Analytic Engine
Development of the first steps for the implementation of privacy preserving pipelines and of a blockchain-based Smart Contract platform
Development of a neural-network-based solution for the provision of privacy to data representations
Extraction of advanced telemetry using P4 programmable switches
Development of a solution for the protection of users against malware and other attacks from the network.
Final definition of MARSAL network components and architecture
Application of reconfigurable functional splits in cell-free networks
Provision of higher level of inter-O- DU operability, of the baseband for MIMO at 60 GHz, and of the mmWave inter-O-DU link
Optimal O-DU cooperation algorithms
Design and validation of Near-RT RIC algorithms
Development of trajectory prediction techniques for dynamic cluster reshaping
Development of data-driven O-RU clustering algorithms
Novel traffic management algorithms that apply load balancing methodologies
Novel ML-based network slicing algorithms
Disaggregation of MEAO’s functionalities
Implementation of privacy preserving data representation algorithms and of the smart contracts platform
Development of policy-driven data integrity and privacy protocols
Development of the P4 assisted network security approach
Integration and testing of the MARSAL software components
Demonstration of MARSAL’s two PoCs and all related scenarios

Impacts:
MARSAL’s innovation aims at providing “smart connectivity technologies for platforms integrating ubiquitous connectivity, storage, and computing resources opening for new service and business models”, “smart connectivity platforms integrating technologies and architectures towards perceived zero latency”, “network scalability towards a high number of resource-constrained IoT devices, multiplicity of service requirements, and new user-controlled connectivity paradigms”, “characterisation and availability of secure and trusted environments for software based virtualised networks, including underlying hardware limitations and enabling trusted multi-tenancy”, “innovative radio spectrum use, novel strategies for coverage/service extension, support of novel wireless technologies and use cases through platforms, usability of today unexplored spectrum”, “dynamic scalability of network capabilities through availability of managed and enhanced optical resources”, “characterisation of ML and blockchain technologies in the connectivity domain, notably for network/service management and security”, and “significant reduction of total cost of ownership through improved operational and capital expenditure efficiency, and energy consumption”. The aforementioned impacts are achieved through the MARSAL innovation in its three pillars, while their effectiveness will be shown through the implementation of the 2 MARSAL POCs.
MARSAL’s societal impact is to contribute to the provision of the 6G network deployment in Europe, through its capabilities to provide new innovative services that demand 6G capabilities. This novel overall approach will also enable telecom and service providers to vision network configurations towards the 6G era that are able to optimize their resource usage thus reducing both CAPEX and OPEX, while also offering secured 6G services at lower costs to consumers.

marsal-network-architecture-4.jpg