PHYSICS · Project

One Platform to Run Serverless Apps Across Any Cloud, Anywhere

digitalPilotedTRL 7

Imagine you run apps in the cloud but you're locked into one provider — switching is painful and expensive. PHYSICS built a kind of universal remote control for serverless computing: you design your app once, and the platform figures out the best place and time to run each piece — whether that's AWS, Azure, an edge server, or your own data center. It optimizes costs and speed automatically, like a GPS that reroutes your data traffic in real time. They tested it on real healthcare, farming, and factory applications across 17 organizations in 9 countries.

By the numbers

consortium partners

countries represented

82%

industry participation ratio

real-world application domains validated

SMEs in consortium

total deliverables produced

demo prototype deliverables

industry partners

The business problem

What needed solving

Companies running serverless applications in the cloud face vendor lock-in, unpredictable costs, and poor performance when workloads span multiple locations or providers. There is no simple way to design a cloud application once and have it automatically run in the optimal place at the optimal time — whether that is on-premise, at the edge, or across different cloud providers.

The solution

What was built

The team built a complete multi-cloud serverless platform with 4 key components: a visual Cloud Design Environment for building applications, an optimized FaaS middleware for cross-site deployment and orchestration, a Backend Optimization Toolkit solving cold-start and multi-tenant problems, and an Artefacts Marketplace (RAMP) for sharing reusable functions and workflows. Both V1 and V2 integrated prototypes were delivered and tested in eHealth, agriculture, and manufacturing applications.

Audience

Who needs this

Cloud service providers looking to offer multi-cloud serverless computingHealthcare IT companies needing data-residency-aware cloud deploymentManufacturing firms running Industry 4.0 workloads across edge and cloudAgriTech companies processing sensor data across distributed infrastructureManaged service providers wanting to reduce client cloud vendor lock-in

Business applications

Who can put this to work

Cloud & IT Services

enterprise

Target: Cloud service providers and managed service providers

If you are a cloud service provider struggling to offer competitive serverless computing without getting locked into a single hyperscaler — this project developed a multi-cloud FaaS platform with deployment optimization and runtime orchestration, validated across 3 real-world applications. The platform handles cold-start problems and multi-tenant interference automatically, tested by a consortium of 14 industry partners.

Healthcare IT

mid-size

Target: Digital health platform operators and hospital IT departments

If you are a healthcare IT company dealing with unpredictable workloads and strict data residency requirements — this project built and tested an eHealth application prototype that runs serverless across multiple cloud locations. The platform decides where to execute based on performance and compliance constraints, with both V1 and V2 prototypes delivered and validated.

Smart Manufacturing

any

Target: Manufacturing companies with Industry 4.0 digital infrastructure

If you are a manufacturer running cloud-based quality control or predictive maintenance but facing high cloud costs and latency issues — this project validated a manufacturing application that optimizes where and when compute tasks run across edge and cloud resources. The consortium included 5 SMEs alongside major industry players, ensuring the solution works at different scales.

Frequently asked

Quick answers

What would it cost to adopt this platform?

The project did not publish pricing or licensing fees. However, the platform is designed to reduce cloud costs by optimizing where serverless functions execute across providers and locations. The RAMP marketplace model suggests a freemium or artifact-sharing approach may be part of the business model.

Can this work at industrial scale with real production workloads?

Yes — the platform was validated in 3 real-world application domains (eHealth, smart agriculture, and smart manufacturing) with both V1 and V2 prototypes delivered. The consortium of 17 partners across 9 countries, with 82% industry participation, indicates the solution was tested under realistic conditions.

Who owns the intellectual property and how is it licensed?

IP is shared among the 17 consortium partners under the Horizon 2020 grant agreement. The project committed to contributing to open source tools and initiatives including Gaia-X and EOSC. Specific licensing terms would need to be discussed with the coordinator GFT Italia SRL.

Does this comply with EU data regulations like GDPR?

The platform's core design optimizes the location of execution, which directly supports data residency requirements. The project aligns with European Strategy for Data and Gaia-X principles. Specific GDPR compliance details would need to be confirmed with the consortium.

How long would integration take for our existing cloud setup?

The Cloud Design Environment enables visual workflow design and integration with existing application components through generalized cloud design patterns. Both V1 (month 15) and V2 (month 32) integrated prototypes were delivered, suggesting a maturing integration process. Exact timelines depend on your current infrastructure.

What cloud providers and platforms does this support?

The platform is designed for multi-cloud and multi-site operation, covering edge resources, multiple cloud providers, and different hardware types. It includes a cross-site FaaS middleware that cooperates with provider-local policies, meaning it adapts to whichever infrastructure you already use.

Is there ongoing support or a community around this?

The RAMP marketplace was built for internal and external contributors to share reusable artifacts like functions, flows, and controllers. The project also contributed to open source initiatives. Post-project support would need to be confirmed with GFT Italia SRL or through the project website at physics-faas.eu.

Consortium

Who built it

This is a heavily industry-driven consortium: 14 out of 17 partners (82%) come from industry, with only 2 universities and 1 research organization. That's unusually high for an EU research project and signals strong commercial intent. Coordinated by GFT Italia, a well-known IT services company, the consortium spans 9 countries across Europe and Israel. The 5 SMEs bring agility while larger industry partners ensure the solution meets enterprise-grade requirements. The 3-year project ran from 2021 to 2023 and produced 27 deliverables including multiple prototype iterations, demonstrating serious execution capability.

GFT ITALIA SRLCoordinator · IT
FSAS TECHNOLOGIES GMBHthirdparty · DE
FUJITSU SERVICES GMBHparticipant · DE
CHAROKOPEIO PANEPISTIMIOparticipant · EL
RYAX TECHNOLOGIESparticipant · FR
CYBELETECHparticipant · FR
DEUTSCHES FORSCHUNGSZENTRUM FUR KUNSTLICHE INTELLIGENZ GMBHparticipant · DE
ATOS SPAIN SAthirdparty · ES
HEWLETT PACKARD ITALIANA SRLparticipant · IT
INNOV-ACTS LIMITEDparticipant · CY
BYTE COMPUTER ANONYMI VIOMICHANIKIEMPORIKI ETAIREIAparticipant · EL
RED HAT ISRAEL LTDparticipant · IL
INNOVATION SPRINTparticipant · BE
ATOS IT SOLUTIONS AND SERVICES IBERIA SLparticipant · ES
UNIVERSIDAD POLITECNICA DE MADRIDparticipant · ES
INQBIT INNOVATIONS SRLparticipant · RO

How to reach the team

GFT Italia SRL (Italy) — a major IT consulting and software company. Contact their innovation or cloud services division.

Order a report on this consortium See GFT ITALIA SRL profile →

Next steps

Talk to the team behind this work.

Want an introduction to the PHYSICS team to discuss licensing their multi-cloud serverless platform for your infrastructure? SciTransfer can arrange a direct meeting with the right technical leads.

Order a report on this topic More in Digital & ICT