SciTransfer
Vitamin-V · Project

Open-Source Cloud Infrastructure Software for RISC-V Processors

digitalTestedTRL 5

Imagine trying to run a modern city's power grid on a set of blueprints that aren't fully finished. This project builds the missing digital tools and manuals so that open-source computer chips can run massive cloud services just as fast as the expensive, proprietary ones we use today. It's like creating a universal translator and a high-tech simulator to ensure everything works perfectly before the hardware is even built.

By the numbers
10
Consortium partners
40%
Industry ratio
19
Total deliverables
The business problem

What needed solving

Cloud providers are dependent on proprietary x86 hardware, leading to high costs and lack of flexibility. Current open-source RISC-V environments lack the virtualization and cryptography features needed to match this performance.

The solution

What was built

A complete RISC-V open-source software stack and a virtual execution environment (VRISC-V) including compilers and toolchains for cloud-specific ISA extensions.

Audience

Who needs this

Cloud Infrastructure ProvidersHardware Architecture DesignersOpen-source OS DevelopersEnterprise AI Deployment Teams
Business applications

Who can put this to work

Cloud Computing
enterprise
Target: Cloud Infrastructure Provider

If you are a provider dealing with high licensing costs and vendor lock-in from x86 hardware — this project developed an open-source software stack that provides iso-performance to x86 counterparts. This allows you to migrate to RISC-V architectures without losing speed. It supports modern setups like Kubernetes and serverless environments.

Cybersecurity
SME
Target: Hardware Security Auditor

If you are a security firm dealing with trust issues in proprietary cloud CPUs — this project developed a toolset for the validation, verification, and testing of software trustworthiness. This ensures that the cloud layers running on RISC-V are secure and free of anomalies. It uses virtual environments like QEMU and gem5 for testing.

AI & Big Data
mid-size
Target: AI Model Deployment Firm

If you are a firm dealing with inefficient AI scaling on standard cloud hardware — this project developed support for vectorization and accelerators for memory compression. This enables the efficient running of AI models like Google Net and VGG19 on open-source hardware. It ensures high-performance execution for BigData tools like Apache Spark.

Frequently asked

Quick answers

What is the cost or price of this software?

Based on available project data, the software stack is described as open-source, implying it is developed for public availability rather than as a priced commercial product.

Can this be deployed at an industrial scale?

Yes, the project specifically targets cloud-scale deployments by porting industry-standard tools like OpenStack, Kubernetes, and Docker to the RISC-V architecture.

What are the IP and licensing terms?

The project focuses on creating an open-source software stack, though specific license types (e.g., Apache or MIT) are not detailed in the provided text.

How does this integrate with existing cloud tools?

It integrates by porting existing suites such as KVM, QEMU, and Kata containers, as well as libraries like JVM and Python, to ensure compatibility with current cloud workflows.

What is the timeline for availability?

The project period is from 2023-01-01 to 2025-12-31, suggesting the final deliverables will be ready by the end of 2025.

Consortium

Who built it

The consortium is well-balanced for technology transfer, consisting of 10 partners across 5 countries. With a 40% industry ratio (4 companies, 4 of which are SMEs), there is a strong link between academic research (3 universities, 3 research centers) and commercial application, ensuring the software stack meets real-world cloud requirements.

How to reach the team

Contact Universitat Politècnica de Catalunya

Next steps

Talk to the team behind this work.

Contact us to explore integration of RISC-V cloud stacks into your infrastructure.