MARSAL · Project

Machine Learning Tools That Make 5G Networks Cheaper and Easier to Manage

digitalTestedTRL 5

Imagine running a highway system where thousands of cars, trucks, and buses all share the same roads, and each type needs different speed limits, lane widths, and safety rules — all managed in real time. That's basically what 5G networks face: tons of different services fighting for the same infrastructure. MARSAL built smart software that uses machine learning to automatically juggle all those network resources, while also making the wireless coverage denser and cheaper through a "cell-free" design where many small antennas work together like a team instead of each covering its own zone. On top of that, they added security layers using AI and blockchain so that different companies sharing the same network can't peek at each other's data.

By the numbers

consortium partners

countries represented

71%

industry partner ratio

total deliverables completed

demonstration deliverables validated

SMEs in the consortium

The business problem

What needed solving

Managing 5G networks is becoming impossibly complex. Operators must juggle dozens of different service types — from autonomous vehicles to factory robots to streaming video — all running on the same shared infrastructure, each with wildly different speed, latency, and reliability requirements. Current manual and rule-based approaches cannot keep up, leading to wasted capacity, poor service quality, and security gaps when multiple tenants share the same network.

The solution

What was built

The project delivered 35 outputs including ML-based algorithms for automatic network resource management, a cell-free wireless networking system demonstrated in dense hotspot areas, converged optical-wireless fronthaul/midhaul architecture, a Virtual Elastic DataCenter management system, and AI plus blockchain-based security for multi-tenant network slicing. Two formal demonstrations validated the cell-free networking and the cognitive assistance security capabilities.

Audience

Who needs this

Mobile network operators deploying or optimizing 5G infrastructureEdge computing and data center providers expanding into telecomSmart venue operators (stadiums, airports, convention centers) needing dense wireless coverageNetwork equipment manufacturers building O-RAN compatible productsEnterprise IT departments planning private 5G networks

Business applications

Who can put this to work

Telecommunications

enterprise

Target: Mobile network operators managing 5G infrastructure

If you are a telecom operator struggling to efficiently allocate network resources across dozens of different service types — this project developed ML-based tools for automatic orchestration of both computing and communication resources. The cell-free networking approach demonstrated in dense hotspot areas could let you serve more users with fewer base stations. With 14 consortium partners including 10 industry players, the technology was validated across real-world conditions.

Cloud and Edge Computing

enterprise

Target: Data center and edge computing providers

If you are a cloud provider looking to extend your services closer to the network edge for 5G applications — this project built a Virtual Elastic DataCenter system that dynamically manages computational resources at both edge and midhaul locations. The ML-based algorithms optimize workload placement across distributed infrastructure. The security mechanisms developed ensure multi-tenant isolation, so multiple clients can safely share your infrastructure.

Smart Venues and Dense Environments

mid-size

Target: Stadium, airport, or convention center operators

If you are managing connectivity in places where thousands of people crowd together — like stadiums, airports, or trade fairs — this project demonstrated cell-free networking specifically for dense and ultra-dense hotspot areas. Instead of relying on a few overloaded cell towers, many small access points cooperate to spread the load evenly. This was validated through 2 dedicated demonstration deliverables focused on exactly these high-density scenarios.

Frequently asked

Quick answers

What would it cost to adopt this technology?

The project was funded as a Research and Innovation Action (RIA), meaning this is pre-commercial technology. Budget details are not available in the dataset. Adoption costs would depend on licensing terms negotiated with the consortium's 14 partners, particularly the coordinator IQUADRAT INFORMATICA SL.

Can this work at industrial scale in real networks?

The consortium demonstrated cell-free networking in dense and ultra-dense hotspot areas, which is a strong signal for scalability. With 10 industry partners out of 14 total (71% industry ratio), the technology was developed with real-world deployment in mind. However, as an RIA project, full commercial-scale deployment would still require additional engineering.

What about intellectual property and licensing?

IP is distributed across 14 partners in 9 countries. The coordinator IQUADRAT INFORMATICA SL (Spain, SME) would be the first point of contact for licensing discussions. As an EU-funded RIA project, results are typically available for licensing, though terms vary by partner.

How does this integrate with existing 5G infrastructure?

MARSAL was designed to contribute to the O-RAN project, which is the open standard most operators are adopting. The converged optical-wireless approach works with existing fronthaul and midhaul segments. This standards-alignment makes integration significantly easier than proprietary alternatives.

Is this technology secure enough for multi-tenant use?

Security was a core focus. The project developed AI and blockchain-based mechanisms specifically for secured multi-tenant network slicing. One of the 2 demonstration deliverables was dedicated to cognitive assistance and its security and privacy implications. This means multiple companies can share the same 5G infrastructure without compromising each other's data.

What is the timeline to market readiness?

The project ran from January 2021 to June 2024 and is now closed. With 35 total deliverables completed including 2 demonstrations, the core technology is validated. Moving to commercial products would likely require 1-2 additional years of productization, depending on the specific component.

Consortium

Who built it

MARSAL assembled a strong, industry-heavy consortium of 14 partners across 9 European countries plus Israel. With 10 industry partners (71% of the consortium) and 4 SMEs, this was clearly built to produce results that work outside the lab. The coordinator, IQUADRAT INFORMATICA SL, is a Spanish SME — meaning the project was led by a company with commercial instincts rather than a purely academic mindset. The geographic spread (Belgium, Cyprus, Germany, Greece, Spain, France, Israel, Italy, Poland) covers major European telecom markets. The 2 university and 2 research institute partners provided scientific depth without the consortium becoming an academic exercise.

IQUADRAT INFORMATICA SLCoordinator · ES
PIETRZYK SLAWOMIRparticipant · PL
ORANGE SAparticipant · FR
CENTRE TECNOLOGIC DE TELECOMUNICACIONS DE CATALUNYAparticipant · ES
INTRACOM SINGLE MEMBER SA TELECOM SOLUTIONSparticipant · EL
ORGANISMOS TILEPIKOINONION TIS ELLADOS OTE AEparticipant · EL
EBOS TECHNOLOGIES LIMITEDparticipant · CY
ACCELLERANparticipant · BE
MELLANOX TECHNOLOGIES LTD - MLNXparticipant · IL
NEC LABORATORIES EUROPE GMBHparticipant · DE
UNIVERSITA DEGLI STUDI DI MILANOparticipant · IT
KATHOLIEKE UNIVERSITEIT LEUVENparticipant · BE
EREVNITIKO PANEPISTIMIAKO INSTITOUTO SYSTIMATON EPIKOINONION KAI YPOLOGISTONparticipant · EL

How to reach the team

IQUADRAT INFORMATICA SL (Spain, SME) — use SciTransfer's coordinator lookup service to find the right contact person

Order a report on this consortium See IQUADRAT INFORMATICA SL profile →

Next steps

Talk to the team behind this work.

Want to explore how MARSAL's ML-based 5G management tools could solve your network challenges? SciTransfer can connect you directly with the right consortium partner for your use case.

Order a report on this topic More in Digital & ICT