Importance of cloud continuum service orchestration in industrial environments

In industrial environments, cloud services and the Internet of Things (IoT) converge, with the need to manage complex and diverse computing environments for service and infrastructure providers, a process called cloud continuum. In this aspect, large-scale cloud data centres require advanced orchestration systems to handle resource allocation for both the massive data generated by IoT devices and the cloud services themselves. This orchestration system requires real-time awareness of all available resources and their status to dynamically allocate resources and enable rapid application deployment.

Therefore, the cloud continuum extends beyond traditional cloud computing by incorporating energy-efficient, low-latency devices closer to data sources at the network edge. This includes extreme edge, edge, and cloud, enabling analysis, processing, storage, and data generation capabilities. Essentially, it connects cloud environments with these distributed devices that act as real-time data sources. Processing this data requires orchestration across various resources scattered throughout the continuum.

This distributed landscape presents unique challenges for orchestration solutions:

• Heterogeneity: Resources are diverse. Spread across layers (cloud, edge, extreme edge), they possess different architectures, operating systems, and capabilities. Orchestration needs to provide seamless access to this complex and varied resource pool.
• Volatility and Mobility: Edge nodes may be unreliable, shutting down, losing connectivity, or changing location. Orchestration must adapt to resource fluctuations and changing environments.
• Scalability: Orchestration needs to handle the massive scale of the continuum, potentially gathering resources from various cloud and edge providers to meet diverse application needs.
• Observability: Efficient mechanisms are needed to collect data on workload status and resource usage across the entire continuum to ensure optimal performance.
• Federated Coordination: Facilitating end-to-end services across cloud, fog, and edge providers requires standardized APIs and interfaces. Orchestration needs to work across administrative domains.
• Policy-Based Management: Efficient runtime mechanisms must enforce policy-based deployment and reconfiguration of applications to meet Service Level Agreements (SLAs) for resource discovery, placement, usage, data processing, and security.
• Security: The orchestration system must guarantee overall system security against various attacks while minimizing the need for user-supplied configurations. This is particularly challenging due to the inherent heterogeneity of resources and potentially separate administrative domains.

6GSMART-ICC provides a container-based orchestration solution supported by Kubernetes across this cloud-to-extreme Edge continuum.

Author: Jesús Benedicto – IT Project Manager & SW Architect at Eviden