Fortissimo: Cloud-based High Performance Modelling sets a new trend
Fortissimo Marketplace publishes new success-stories of innovative small and medium sized enterprises
- Successful start of Cloud-based HPC-platform Fortissimo Marketplace
- New Success Stories demonstrate potential of HPC-as-a-Service
- Forecast for 2017 shows strong demand for HPC-services
Following its official launch on October 31 2016, Fortissimo Marketplace has garnered strong interest from its target group. The Cloud-based platform simplifies access for small and medium sized enterprises (SME) to high performance computers. The approach using Cloud-services can cut down complex and expensive setup costs for High Performance Modelling and enables flexible accounting of service usage. Service providers benefit from an accurately targeted sales channel, perfect alignment with client requirements and integrated billing services.
High Performance Computing (HPC) and the simulation and modeling software run on those systems are widely seen as an effective but very expensive design and development tool. The initial cost for setup and operation can be quite high and intimidate potential users, particularly small and medium sized companies. The Cloud-based portal Fortissimo Marketplace employs a self-service approach to reduce costs significantly and permits end users to choose the best combination among HPC providers, software manufacturers and technology experts. Professor Mark Parsons, CEO Fortissimo Marketplace Ltd., sees a lot of potential for the next year: „Throughout the Fortissimo project we’ve seen a significant increase in smaller businesses harnessing the availability of on demand low cost super computing power to model their business processes and products in ways that have previously only been possible for large enterprises. The availability of on demand supercomputing is a transformative technology that can help SMEs compete on a level playing field with large enterprises.“
Cloud-migration shortens development cycles, optimizes profit
Recently published success stories demonstrate the broad application of HPC-as-a-service. Spanish SME Matrici, a manufacturer of very complex stamped sheet metal parts for the automobile and aerospace industry, was able to reduce its costs by about 1.5 Million Euro annually after moving to a Cloud-based variant of the CAM-software Stampack from Quantech. Stampack simulates the metal stamping process, so that the construction and production process runs as smoothly as possible. However, the software requires a lot of processing power which a small SME like Matrici has no cost-effective way of providing. After Quantech had ported its software for use in a Cloud-based environment, Matrici was able to complete design processes in half the time formerly needed, thereby cutting development costs by half. HPC provider CESGA and Fraunhofer SCAI supported the migration of Stampack to StamHPC and expect interesting business benefits as well. A projected number of 300 SMEs are expected to become new StamHPC users within the next five years, generating a revenue of roughly 60 Million Euro.
HPC usage without the steep learning curve
Another resounding success can be reported from chimney specialist manufacturer DINAK. This highly innovative business was looking for a way to virtualize very cumbersome tests for the optimization of chimney construction. So far, DINAK was only able to address this challenge with intensive experimental tests, rebuilding elements of the product in the lab and measuring air flow, temperatures and isolation properties. After HPC experts Aimen, in collaboration with HPC provider CESGA, successfully developed a web-based tool, using the Open Source HPC software OpenFOAM, DINAK engineers were able to virtualize the optimization process. Making the solution user-friendly was a key part of the project. This way, even engineers with little or no experience of using Computational Fluid Dynamics (CFD) were able to use the software quickly and efficiently. Another boon was the remote visualization concept, used in the web app. This way, no huge data files needed to be downloaded to the client’s computers, which would have required very high-bandwidth internet access.
New business model through cloud-based HPC application
Another interesting application is Oxolutia’s Cloud-based simulation of high-temperature superconducting cables. The Spanish SME develops, among other materials, superconductors that gain their highly conductive characteristics at temperatures (> -196°C) well above absolute zero. This way, transmissions are almost completely loss-free and extremely efficient. HPC applications are an indispensable tool for optimising development cycles. In order to utilize Cloud-based HPC resources, a new application, based on the well-known Fempar-software was developed. It can simulate the magnetic, thermal and electrical parameters of superconductors, as a result reducing development costs of prototypes and decreasing overall costs dramatically. This way, for the first time, SMEs are able to take advantage of these sophisticated tools. The application is now offered as a pay-per-use-service, complete with the necessary expertise and support. Oxolutia estimates that during the next three years, the software will generate additional revenue in the excess of 500.000 Euros.
Fortissimo Marketplace generates possibilities for users and service providers
The goal of the Fortissimo Marketplace is to provide SMEs with cost-effective access to advanced simulation and modeling services through a Cloud infrastructure of HPC resources, software applications, expertise, and tools. Larger enterprises or academic institutions can of course benefit from the advanced simulation and modeling services as well. The Marketplace is open to any service provider and offers flexible terms for pricing, payment and service presentation. Clients can use these services immediately after registering on the platform, which is free of charge. All Fortissimo success stories are available by following this link: https://www.fortissimo-project.eu/experiments.
Share it: