Main Technical Results Aligned with Project Objectives

1. Deployment of an Elastic Edge Computing architecture (Objective 1)
   ▪ Defined and implemented a cloud‑native Elastic Edge Computing (EEC) paradigm that overcomes resource fragmentation and under‑utilization at the network edge.
   ▪ Extended the MEC system to support both vertical and horizontal disaggregation of containerized applications across multiple Edge nodes.
   ▪ Achieved near‑zero perceived latency for critical services by dynamically balancing workloads and latency budgets across Edge and core.

2. Autonomous, distributed control with embedded Machine Learning (Objective 2)
   ▪ Delivered a hierarchical orchestration framework with Analytic Engines at all Edge levels and Decision Engines at core orchestrator subsystems (NFVO and multi‑access Edge orchestrator).
   ▪ Integrated Machine Learning models such as Reinforcement Learning and Embedding Propagation to generate context representations and drive proactive resource reconfiguration.
   ▪ Enabled closed‑loop autonomy across all infrastructure layers, avoiding the scalability limitations associated with centralized control.

3. Implementation of zero‑touch network slice reconfiguration (Objective 3)
   ▪ Developed mechanisms for automatically reconfiguring network slices in real time based on changing traffic and service demands.
   ▪ Validated adaptive slice management in realistic environments, demonstrating automated allocation of compute, storage, communications, and core resources.
   ▪ Confirmed compliance with strict 6G requirements through vertical use cases such as automotive object identification and collision detection.

4. Dynamic Virtual Network Embedding and multi‑objective resource optimization (Objectives 1 & 2)
   ▪ Studied and deployed algorithms for Dynamic Virtual Network Embedding to optimally place MEC application functions under compute, network, storage, and latency constraints.
   ▪ Integrated predictive analytics that continuously processed telemetry data to anticipate resource demand and trigger proactive adjustments.
   ▪ Enabled efficient utilization of heterogeneous resources including general‑purpose CPUs and FPGA accelerators via hardware abstraction layers and partial reconfiguration.

5. Cloud‑native MEC application integration and traffic steering (Objectives 1 & 3)
   ▪ Enabled deployment of MEC applications as cloud‑native artifacts (Helm charts) with support for both horizontal and vertical scaling.
   ▪ Implemented traffic steering mechanisms that dynamically redirect user traffic to optimal MEC hosts for load balancing and SLA fulfillment.
   ▪ Ensured seamless orchestration of MEC functions in coordination with 5G core components (UPF, AMF, SMF, PCF) to support end‑to‑end service delivery.

6. End‑to‑end Proof of Concept integration and validation (Objective 3)
   ▪ Integrated all architectural components and enablers into a prototype architecture validated under real traffic conditions in the Castellolí testbed.
   ▪ Verified functional and non‑functional interoperability of orchestration engines, analytics subsystems, MEC platform, and 5G network functions.
   ▪ Demonstrated proactive zero‑touch slice reconfiguration to maintain performance under variable load, confirming feasibility for future 6G service scenarios.

7. Dissemination, exploitation, and ecosystem impact (Objective 4)
   ▪ Published project results in high‑impact journals and premier IEEE conferences, enhancing scientific visibility and influence.
   ▪ Presented innovations at global industry events, including MWC 2025, reinforcing the position of Spanish R&I in next‑generation networks.
   ▪ Strengthened industrial partnerships and enabled further R&I proposals (e.g., SNS JU project VERGE), expanding potential adoption in telecom and cross‑industry verticals such as automotive, energy, and extended‑reality media.