Concept
ELASTIC aims to advance 6G service orchestration by leveraging cutting-edge cloud-native technologies. By incorporating WebAssembly, serverless FaaS, TEE-enabled Confidential Computing, and eBPF/XDP, ELASTIC addresses the limitations of existing in-network and edge computing systems. These technologies collectively enable ELASTIC to efficiently manage a diverse range of workloads, including complex machine learning tasks and dynamic security policies, ensuring robust performance and adaptability in modern 6G environments.
The framework provides a comprehensive solution for orchestrating data and services securely and efficiently. ELASTIC supports privacy-preserving multi-party AI and federated machine learning, ensuring data authenticity and trusted interactions across dynamic service contexts. With its high level of abstraction and programmability, ELASTIC utilizes node containers accessed through a hardware abstraction layer to facilitate seamless cloud-to-edge deployment. This approach ensures that ELASTIC is capable of managing complex services across the entire network architecture, from core infrastructure to edge devices, offering a scalable and versatile solution for 6G service orchestration.
Current TRL and Expected Advancements
Wasm-Operator SDK and runtime for compiling, packaging, and running serverless Kubernetes operators as Wasm modules (IMEC)
- TRL Start-End: 3-4
- Remarks: A framework that recompiles Kubernetes agents into a serverless control plane with minimal code changes, including a Wasm runtime for serverless Kubernetes operators and features predictive loading/unloading of controllers to balance responsiveness and resource usage.
WasmHAL SDK interfaces and runtime extensions for securely connecting Wasm applications to hardware across platforms (IMEC)
- TRL Start-End: 2-4
- Remarks: An SDK for compiling hardware drivers and hardware abstraction layers to Wasm, providing interfaces for application connection to packaged HAL and runtime extensions that enable curated access to underlying hardware through a WebAssembly runtime.
Automatic MAC profiles for Wasm runtime containers (LUN)
- TRL Start-End: 3-4
- Remarks: Creation of access profiles for Wasm runtime containers, tailored for typical WebAssembly classes running on Docker. The profile generator will be integrated with the WebAssembly test and integration environment.
Automation tooling for confidential computing environments (LUN)
- TRL Start-End: 3-4
- Remarks: Development of tools for secure execution, application updates, deployment monitoring, and migration.
Automatic MAC profiles for Wasm runtime containers (LUN)
- TRL Start-End: 3-4
- Remarks: Creation of access profiles for Wasm runtime containers, tailored for typical WebAssembly classes running on Docker. The profile generator will be integrated with the WebAssembly test and integration environment.
Automation tooling for confidential computing environments (LUN)
- TRL Start-End: 3-4
- Remarks: Development of tools for secure execution, application updates, deployment monitoring, and migration.
Mobility attack robust IoT resource allocation model (LUN)
- TRL Start-End: 2-3
- Remarks: Development of new resource allocation models that are robust against malicious mobility patterns intended to influence NWDAF-based resource allocations.
Static eBPF code security analyzer (POLITO)
- TRL Start-End: 2-3
- Remarks: A static analyzer capable of identifying security vulnerabilities in eBPF code with reference to the source code.
Polycube framework (POLITO)
- TRL Start-End: 3-4
- Remarks: A framework to create and deploy eBPF programs.
AI-based Intrusion Detection/Prevention System (TUC)
- TRL Start-End: 3-4
- Remarks: An AI-based Intrusion Detection/Prevention System mapped on reconfigurable platforms, offering online ML training and applying trained models on streaming data. The goal is to offload cluster security tools from real-time network monitoring of security issues.
Hardware-based cryptography module (TUC)
- TRL Start-End: 2-3
- Remarks: Integration of hardware systems like GPUs and FPGAs into the ELASTIC structure, applying cryptography and certification keys on internal cluster moving data based n the cluster security status and user needs.
Federated Learning as a Service (FLaaS) (TID)
- TRL Start-End: 4-6
- Remarks: Development of the first Federated Learning as a Service (FLaaS) platform, allowing third-party applications and services to build FL models in a seamless and transparent fashion on user devices, supporting higher resource heterogeneity, hierarchical layers of networks and devices, and more privacy-preserving ML models