MIKELANGELO · Project

Faster Cloud Computing That Cuts Virtualization Overhead for Simulations and Big Data

digitalTestedTRL 6

When you run software in the cloud, there's a hidden tax — layers of operating system and virtualization code sit between your application and the actual hardware, slowing everything down. MIKELANGELO stripped away that bloat by building a super-lean operating system (OSv) and a faster version of the standard KVM hypervisor. Think of it like replacing a heavy SUV with a go-kart for a warehouse job — same work, fraction of the weight. They proved it works on real industrial tasks like aerodynamic simulations, big data processing, and even bone tissue modeling.

By the numbers

EUR 5,993,419

EU contribution for virtualization R&D

consortium partners across 4 countries

real-world use cases tested and validated

total project deliverables produced

67%

industry partners in consortium

SMEs in the consortium

The business problem

What needed solving

Companies running heavy computational workloads in the cloud — simulations, big data analytics, HPC jobs — lose significant performance to virtualization overhead. The layers of software between the application and the hardware (hypervisor, guest OS, virtual I/O) add latency and waste resources, making cloud-based HPC slower and more expensive than bare-metal alternatives.

The solution

What was built

The project built an integrated virtualization stack combining a lightweight guest operating system (OSv), an optimized hypervisor (SuperKVM with improved virtual I/O), and packaging tools for OpenStack and OpenNebula deployment. This was validated through 4 use cases: cloud bursting, big data (at GWDG), aerodynamic simulations (OpenFOAM), and bone tissue simulation — producing 33 deliverables in total.

Audience

Who needs this

Simulation-dependent SMEs running CFD or FEA in the cloud (e.g., OpenFOAM users)Cloud service providers looking to reduce virtualization overhead and improve marginsHPC centers transitioning to cloud-based service delivery modelsBig data analytics firms needing faster I/O throughput in virtualized environmentsAutomotive and aerospace companies outsourcing aerodynamic computations to cloud HPC

Business applications

Who can put this to work

Engineering & Simulation Services

SME

Target: SMEs running computational fluid dynamics or finite element simulations

If you are a simulation-dependent SME paying for cloud HPC time to run tools like OpenFOAM — this project developed a lightweight virtualization stack (OSv + SuperKVM) that reduces I/O overhead in cloud environments. Their aerodynamic map use case specifically tested deployment of OpenFOAM cases and evaluated the performance benefits for industrial aerodynamic calculations. Fewer wasted CPU cycles means faster results and lower cloud bills.

Cloud & Data Center Operations

enterprise

Target: Cloud service providers and managed hosting companies

If you are a cloud provider struggling with virtualization overhead eating into your margins — MIKELANGELO improved the core KVM hypervisor's I/O performance and packaged it with OpenStack and OpenNebula integration. With 6 industry partners including the original creator of KVM on the team, the project tackled one of the last major performance bottlenecks in virtualization. This means more workloads per physical server and better responsiveness for customers.

Research Computing & HPC

enterprise

Target: HPC centers and universities offering cloud-based computing services

If you are an HPC center looking to cloudify your computing services — MIKELANGELO specifically addressed this with a large HPC centre partner that had the business need to move HPC workloads to cloud infrastructure. Their cloud bursting use case tested strategies for spilling peak workloads into the cloud without sacrificing performance. The big data use case was validated by GWDG's customers, demonstrating real-world applicability.

Frequently asked

Quick answers

What would it cost to adopt this technology?

MIKELANGELO was built on open-source components — OSv and KVM are both open-source projects. The total EU investment was EUR 5,993,419 across 9 partners over 3 years. Adoption costs would primarily involve integration effort and testing with your existing OpenStack or OpenNebula infrastructure, not licensing fees.

Can this scale to production workloads?

The project tested at industrial scale through 4 distinct use cases: cloud bursting, big data processing (validated by GWDG customers), aerodynamic map calculations with OpenFOAM, and cancellous bone simulation. These were real workloads, not lab demos. The architecture integrates with standard cloud platforms (OpenStack, OpenNebula) designed for production scale.

What about intellectual property and licensing?

The core technologies — KVM and OSv — are open-source. The project consortium included 4 SMEs and 6 industry partners who contributed to exploitation planning. Based on available project data, specific IP arrangements would need to be discussed with the coordinator XLAB or individual technology partners.

How does this integrate with existing infrastructure?

MIKELANGELO was designed for heterogeneous infrastructures, packaged under OpenStack or OpenNebula — the two most widely used open-source cloud management platforms. This means it can layer onto existing cloud deployments without requiring a complete infrastructure replacement.

Is this still maintained after the project ended in 2017?

The project closed in December 2017. OSv and KVM continue as active open-source projects with their own communities. Based on available project data, specific MIKELANGELO enhancements may have been merged upstream or maintained by consortium members like XLAB. Current status should be verified with the coordinator.

What industries were actually tested?

Four concrete use cases were implemented and validated: cloud bursting for general cloud workloads, big data processing for data center customers, aerodynamic simulations for automotive/aerospace engineering, and cancellous bone tissue simulation for biomedical research. Each had dedicated deliverables with external end-user validation.

Consortium

Who built it

The MIKELANGELO consortium was heavily industry-driven with 6 out of 9 partners (67%) from industry and 4 being SMEs, spread across 4 countries (Germany, Ireland, Israel, Slovenia). The coordinator XLAB is a Slovenian software SME, which signals a practical, market-oriented project lead rather than a purely academic exercise. Notable is the involvement of the original creator of KVM (Avi Kivity), lending deep technical credibility. The 2 university partners provided the HPC and big data validation environments. This industry-heavy composition suggests the technology was developed with commercial adoption in mind from day one.

XLAB RAZVOJ PROGRAMSKE OPREME IN SVETOVANJE DOOCoordinator · SI
HUAWEI TECHNOLOGIES DUESSELDORF GMBHparticipant · DE
PIPISTREL DOO PODJETJE ZA PROIZVODNJO ZRACNIH PLOVILparticipant · SI
BEN-GURION UNIVERSITY OF THE NEGEVparticipant · IL
INTEL RESEARCH AND DEVELOPMENT IRELAND LIMITEDparticipant · IE
UNIVERSITY OF STUTTGARTparticipant · DE
GESELLSCHAFT FUR WISSENSCHAFTLICHE DATENVERARBEITUNG MBH GOTTINGENparticipant · DE
IBM ISRAEL - SCIENCE AND TECHNOLOGY LTDparticipant · IL

How to reach the team

XLAB (Slovenia) coordinated this project — contact them for licensing, integration support, or access to the MIKELANGELO virtualization stack.

Order a report on this consortium See XLAB RAZVOJ PROGRAMSKE OPREME IN SVETOVANJE DOO profile →

Next steps

Talk to the team behind this work.

Want to know if MIKELANGELO's lightweight virtualization can cut your cloud computing costs? SciTransfer can connect you with the right consortium partner for your use case.

Order a report on this topic More in Digital & ICT