Cognitive Cloud Continuum

The Cognitive Computing Continuum is the concept underlying AVANT’s infrastructural approach.

More than just a network of physical resources, it can be understood as an intelligent and dynamic environment where edge nodes, cloud, and distributed resources collaborate in a flexible and adaptive manner.

The goal is to overcome the current fragmentation of computational infrastructures.

Features

Its main features, in relation to AVANT, are:

  • Dynamic and collaborative integration – The continuum connects edge and cloud nodes from multiple providers into a heterogeneous, self-managed system capable of learning from operational conditions.
  • “Liquefied” computing – AVANT contributes to making computing flexible enough to scale across multi-cloud environments and dynamically adapt to network capabilities.
  • Intelligent orchestration – The cognitive component is provided by the C3OP platform (Cognitive Computing Continuum Orchestration Platform), developed within the project, which optimizes the use of cloud-edge resources.
  • Autonomous adaptation – Unlike approaches focused solely on interoperability (such as Gaia-X), AVANT manages heterogeneous resources in a self-regulating manner, dynamically adapting to system conditions and data topologies.
  • Multi-objective optimization – Orchestration is not limited to resource allocation alone but simultaneously considers availability, load, energy consumption, reputation, and application requirements.

In summary, the Cognitive Computing Continuum acts as a virtual operating system for Digital Twins and other data-intensive applications, allowing them to use distributed resources efficiently, securely, and sustainably through a software infrastructure that combines intelligence, autonomy, and flexibility.

Documents
Document

A Greedy Data-Anchored Placement of Microservices in Federated Clouds

This paper proposes an algorithm that optimizes multi-cloud deployment, improving response time without violating data movement constraints.

Discover more
Document

Data-anchored multi-cloud microservices placement: a greedy approach

This extended abstract proposes the use of the WL-A model with “anchors” to achieve a data-centric deployment, reducing data movement and improving response times.

Discover more