AI-NET ANTILLAS: Automated network telecom infrastructure with intelligent autonomous systems
Infrastructure for Smart and Secure Communication Networks
Communication networks have always been both feats of engineering and proof of our ability to manage complexity. From the first postal system to the manual switchboards featured in many period movies and on to the globe-spanning, dynamic digital networks of today, the challenge has always been to strike the right balance and keep them fit for purpose. Networks need to be robust and reliable, but also pliant and flexible – and above all, they need to be safe and secure.
More and more entities are joining the conversation, not limited to human actors, but including machine-to-machine communication in smart industry or the high-throughput, real-time communication needed for autonomous cars. Novel technologies are creating entire new networks on top of the existing landscape, and trends like cloud or edge computing are shifting the network burden in unexpected ways. Orchestrating this cacophony of infrastructures, applications, data, and users has become impossible without the aid of powerful algorithms. The direction of travel is clear: Towards fully automated networks that can configure and maintain themselves in a fine balance between physical and logical locations, technical reliability, bandwidths and latency, performance and efficiency, and security for both the infrastructure itself and the information travelling through it.
The AI-NET ANTILLAS (Accelerating digital transformation in Europe by Intelligent NETwork automation – Automated Network Telecom Infrastructure with InteLLigent Autonomous Systems) project intends to develop a platform for novel services that will make this vision become a reality. Its overarching goal is to produce applied proof-of-concept test cases for concrete application scenarios in its three subprojects and contributing to the political ambition of promoting society-wide digitalization in a strong partnership between state and private enterprise actors.
Objectives and Approach
The partners pursuing the SEBIU (Safe integration into automated industrial environments) sub-project headed by Creonic / NOKIA are developing the fundamental infrastructure components for automated communications networks in the high-profile fields of digital industry and autonomous driving. Their intention is to allow the end-to-end automation of the physical, network, and service infrastructure to remove the need for manual control and pave the way for truly autonomous network operations, including autonomous configuration, compliance, real-time data collection, data analysis, troubleshooting, software update.
Innovations and Prospects
Drawing directly on the company’s expertise and experience with secure elements and use of cryptographic protection technology in software and cloud scenarios and its recent emphasis on safeguarding machine learning models, Wibu-Systems is using special hardened secure elements to create an environment that operates the proposed autonomous networks with the required software and ML models without fear of external tampering and prepares secure pathways for real-time maintenance and patching to guarantee the active and reliable operation of the network. Particular attention is paid to preparing the security infrastructure to accommodate the provisioning of machine learning models.
The innovations at the network infrastructure level that Wibu-Systems is attending to focus on the following areas:
Flexible and powerful cloud Licensing: Wibu-Systems will evaluate cloud implementations for performance and scalability to ensure that the Wibu-Systems cloud implementation satisfies the requirements of future solutions.
Architectures to increase flexibility and dynamism in deployment and operation: Wibu-Systems will contribute to dynamic update and provisioning mechanisms
Secure integration into automated environments: Wibu-Systems will leverage security measures for network components (e.g. trusted code execution or compartmentalization for 3rd party code) supported by secure hardened elements (Secure Boot, ARM TrustZone, WIBU CloudProtect).
Secure and privacy compliant AI/ML models: Wibu-Systems will protect the entire machine learning lifecycle (integrated data acquisition, training, provisioning and operation of the models, e.g. on an external GPU).
The results will be exemplified in a dedicated demo system for an industrial cloud environment.