The CAPE project develops a pc structure for environment friendly Edge-Cloud.
Regionally deployed highly effective edge-cloud Infrastructure is required to assist AI-driven environments with networks of autonomous units to keep up their context and particular person and shared states. This enables the units to work collectively in federation in the direction of particular person and shared objectives. To scale sustainably, this infrastructure should be offered as a service and constructed on composable {hardware} foundations. Compute, reminiscence, storage, and accelerators should dynamically self-configure to maximise utilisation, minimise waste, and adapt capability and heterogeneity to native wants delivering cloud-class efficiency via decentralised, fine-grained deployment.
The EU-funded CAPE (European Open Compute Structure for Highly effective Edge) venture addresses this hole by establishing edge micro knowledge centres as a brand new, composable constructing block of the Edge–Cloud Continuum. CAPE combines open {hardware}, open-source software program, and open requirements to allow versatile, environment friendly, and sovereign edge computing throughout Europe.
Edge {hardware} platforms: Composable by design
CAPE rethinks edge servers as swimming pools of dynamically composable sources relatively than fixed-function techniques. The platform is designed as autonomous, fine-grained edge micro knowledge centres that may function independently whereas federating with neighbouring nodes to kind a distributed edge-cloud cloth. Compute, reminiscence, storage, and accelerators could be flexibly mixed to match workload necessities, considerably bettering utilisation and decreasing overprovisioning.
The venture develops two complementary open {hardware} platforms primarily based on COM-HPC know-how: the embedded Excessive-Efficiency Server (eHPS) and the Embedded Micro Knowledge Middle (EMDC). These platforms assist heterogeneous compute nodes and goal a variety of edge eventualities, from industrial on-premise deployments to telecom environments.
Composable {hardware} is enabled via PCIe switching mixed with Compute Specific Hyperlink (CXL), which gives cache-coherent connections between processors, accelerators, and pooled reminiscence. This enables sources to be dynamically pooled and reallocated at runtime, enabling every autonomous edge unit to adapt to fluctuating AI workloads whereas maximising {hardware} utilisation.
Power effectivity is a core design goal. By maximising {hardware} utilisation and decreasing stranded sources throughout distributed deployments, CAPE minimises overprovisioning and pointless power consumption. Passive cooling, waste warmth restoration, and building-level integration additional cut back the environmental footprint of embedded edge infrastructure.
Software program stack: Orchestrating the Edge-Cloud Continuum
To completely exploit {hardware} composability, CAPE develops an open-source software program stack that hides infrastructure complexity whereas preserving management and transparency. The stack combines Kubernetes-based multi-cluster orchestration with AI-assisted automation. Orchestration preserves native autonomy, making certain that every edge unit can function independently whereas collaborating in federated coordination throughout the distributed infrastructure.
A central innovation is Infrastructure from Code (IfC), which derives infrastructure definitions instantly from software code, decreasing growth effort in comparison with conventional Infrastructure as Code approaches. Purposes could be deployed seamlessly throughout CAPE edge servers and exterior cloud environments, enabling elastic scaling and workload migration.
Unified administration is offered by openMPMC, an open multi-platform administration controller providing web-based operation and Redfish-compatible APIs for configuration, monitoring, and dynamic useful resource composition.
Standardisation and openness
CAPE is totally dedicated to open requirements and open-source growth. By constructing on applied sciences akin to COM-HPC and CXL and actively contributing to European and worldwide initiatives, the venture avoids proprietary lock-in and fosters a sustainable ecosystem for edge computing. This openness instantly helps Europe’s objectives for digital sovereignty throughout the Edge–Cloud Continuum by way of the Open Federated Edge-Cloud Infrastructure Alliance.
Use circumstances: Influence in real-world environments
CAPE validates its structure via three consultant use circumstances. In sensible grids, edge servers allow real-time monitoring and anomaly detection at substations. For deep studying on the edge, heterogeneous accelerators and CXL-based reminiscence pooling assist environment friendly AI inference underneath strict energy constraints. In satellite tv for pc payload knowledge processing, CAPE allows near-real-time analytics via dynamic useful resource composition throughout edge and cloud.
Outlook
CAPE demonstrates how composability, open requirements, and clever orchestration can remodel edge computing into a versatile and energy-efficient basis for future digital companies. By tightly integrating {hardware} and software program innovation, the venture lays the groundwork for scalable, sovereign edge infrastructure in Europe.
Disclaimer
The CAPE venture has acquired funding from the European
Union’s Horizon Europe analysis and innovation programme underneath grant settlement No 101189899, and was co-funded by the Swiss Nationwide Science Basis (SNSF, No. 200429), and by Armasuisse Science and Know-how.
Please Observe: It is a Industrial Profile
Please notice, this text will even seem within the twenty fifth version of our quarterly publication.
