CloudLightning focused on self-organizing heterogeneous cloud resources (GPU, MIC, DFE), and DICE addressed data-intensive cloud application quality.
INSTITUTUL E-AUSTRIA TIMISOARA
Romanian research centre specializing in heterogeneous cloud computing, software quality modeling, and exascale data processing frameworks.
Their core work
E-Austria Timisoara is a Romanian research centre specializing in cloud computing architectures, high-performance computing, and data-intensive application design. Their work focuses on making heterogeneous computing resources — GPUs, many-integrated-core processors, and data flow engines — work together efficiently in cloud environments. They contribute applied research on software quality, reliability modeling, and programming models for processing large-scale data, bridging the gap between hardware diversity and usable cloud services.
What they specialise in
DICE specifically targeted iterative quality enhancements using UML/MARTE/TOSCA modeling for big data cloud applications.
ASPIDE (2018-2021) addressed programming models for exascale data processing, representing a scale-up from their earlier cloud work.
DICE applied UML, MARTE profiles, and TOSCA standards to model and ensure quality of cloud-deployed applications.
How they've shifted over time
Their early H2020 work (2015-2018) centred on self-organizing heterogeneous cloud infrastructure — orchestrating GPUs, many-integrated-core chips, and data flow engines within cloud platforms. By 2018, the focus shifted toward software quality assurance for cloud applications (using formal modeling standards like UML/MARTE/TOSCA) and then further toward exascale programming models for extreme data processing. The trajectory shows a clear move from infrastructure-level cloud research toward higher-level programming abstractions and ever-larger data scales.
Moving from cloud infrastructure optimization toward programming frameworks for extreme-scale data, suggesting future contributions in HPC and big data middleware.
How they like to work
IEAT operates exclusively as a project participant — they join consortia rather than leading them, which is typical for a specialized research centre contributing domain expertise to larger efforts. With 23 unique partners across 10 countries from just 3 projects, they work in medium-to-large consortia and have broad rather than concentrated partnership patterns. This suggests they are an accessible, reliable partner who integrates well into diverse international teams.
Despite only three projects, IEAT has built a network spanning 23 partners across 10 countries, indicating participation in sizable European consortia with good geographic diversity rather than a narrow regional cluster.
What sets them apart
IEAT sits at the intersection of cloud computing, heterogeneous hardware acceleration, and formal quality modeling — a combination that few research centres in Romania or the wider region offer together. Their progression from GPU/FPGA cloud orchestration to exascale programming models gives them practical experience across the full stack of modern high-performance distributed computing. For consortium builders, they offer a cost-effective Romanian partner with genuine technical depth in cloud and HPC software research.
Highlights from their portfolio
- CloudLightningTackled the complex challenge of self-organizing cloud infrastructure across GPUs, MIC, and data flow engines — a technically ambitious heterogeneous computing project.
- ASPIDEAddressed exascale programming models for extreme data processing, representing IEAT's move into the highest tier of computational scale.