2023-12-002 - MD-NAF - Massively Decentralized Network Automation Framework

MD-NAF

2023-12-002

The proposed Massively Decentralized NAF (MD-NAF) technology creates a new NAF concept where multiple (or every) geo-distributed network locations run a portion of the complete NAF architecture. The introduction of the MF-NAF technology might have a significant impact in Network Automation market, where innovation cycles are slow, due to slow buying cycles of telecom industry and infrastructure operators

Environmental: energy reduction; Social: safer and more resilient networks.

Optimize infrastructure and network operations towards error-free network automation while reducing Total Cost of Ownershipand complying with stringent demands imposed by beyond 5G infrastructures.

Novel 5G and envisioned beyond 5G infrastructures and networks pose a real challenge on how they have to be managed. Extremely fast reaction and prediction times have to be achieved to fulfill the proposed KPIs. Novel network automation frameworks propose running automated closed control loops in a centralized location; however, that prevents achieving low-latency management interactions and blackouts in the event of connectivity issues. The proposed Massively Decentralized Network Automation Framework (MD-NAF) eliminates single points of failure and minimizes infrastructure interaction latencies by deploying MD-NAF--enabled nodes next to the equipment to be controlled.

30000000000

580000000

19570000

OpenSource - Freemium. Patent Licensing

Provide means to efficiently automate the infrastructure and network operation, in an error-free, energy-efficient, resilient, highly available, and secure manner.

- Equipment manufacturers - Network and infrastructure operators

Protection of MD-NAF technology. Future aspects related to MD-NAF might need IPR protection, such as procedures, architectures, workflows, and methodologies.

IPR protection

a) initial IPR protection; b) preliminary product releases; c) preliminary proof-of-concepts at the laboratory; d) find additional funding.

MD-NAF goes one step beyond proposing to manage the 5G and beyond infrastructures using masively geo-distributed NAF instances colocated with the infrastructure nodes enabling high-availability, resiliency, and close-to-real-time automation.

MD-NAF radically changes current paradigm on centralized SDN controllers. Closed market, and missing IPR protection.

2-year 4-workpackage project. WP1: Project Management. WP2: MD-NAF Technology Development. WP3: Market Exploitation. WP4: Product Protection, Regulation, and Dissemination. Development plans: SCRUM agile methodology first appointing a Product Owner, then running 7 3-month SCRUM iterations each consisting in development, testing, validation and release. Laboratory proof-of-concept demonstration will be delayed due to lack of funding. Comercialization strategy will be developed to find customers and other sources of funding to reach higher TRLs.

Patent request submitted. Positive patentability opinion, pending publication.

Patent request sent. Received European Search Report with a quite-positive patentability opinion. Pending to define internal defense strategies. Development slowly progressing due to lack of resources and funding.

The team has large experience in internaitonal, EU, national, and regional R&D projects related to infrastructure and network management. Ricard Vilalta and Lluis Gifre are, respectively, Chair and Technical Chair of the ETSI TeraFlowSDN Source Development Group in charge of a cloud-native micro-service--based SDN controller leveraging network automation features. Ricard Vilalta and Raul Muñoz are co-inventors of parent "METHOD AND SYSTEM FOR CLOUD-NATIVE APPLICATIONS-BASED NETWORK OPERATIONS" (EP3809631A1) describing a cloud-native SDN controller architecture.

CTTC ensures the quality of the results and the efficient management of all the R&D projects it is involved by following the "CTTC R&D project management procedures", certified by AENOR UNE 166002:2021.

The proposed Massively Decentralized NAF (MD-NAF) technology goes one step beyond current state-of-the-art C&M systems and paves the way for network and infrastructure operators worldwide to adopt Network Automation technology empowered with Artificial Intelligence. It focuses on creating a new NAF concept where multiple (or every) geo-distributed network locations run a portion of the complete NAF architecture. This NAF portion is composed of the basic elements to keep the equipment up and running, to enable interoperation with neighbour MD-NAF-enabled locations, and to distribute and run non-critical procedures for scalability purposes. To achieve this, the MD-NAF technology addresses four main problems: (i) the partial data replication and smart distribution of replicas among the available network locations, (ii) the definition of the architectural deployment per network location to achieve the MD-NAF solution, (iii) the negotiation protocols between the reachable MD-NAF instances to maintain subsets of the NAF operational, and (iv) the self-reconciliation and re-optimization of the NAF instances in the event of a disaster recovery.