eccenca leading the charge for Next-Gen Edge Computing

Since January 2023, eccenca has been collaborating with esteemed European partners to develop middleware components that enable Edge-AI. Edge AI refers to the practice of deploying artificial intelligence algorithms and models directly on local devices, known as edge devices. These devices, such as sensors, smartphones, and IoT devices, process data locally rather than sending it to a centralized cloud server. This approach enables real-time data analysis and decision-making at the location where the data is generated.

The work is part of the CLEVER Collaborative edge-cloud continuum and embedded AI for a Visionary industry of the future, a project founded by the European Commission that aims to revolutionize edge computing by enhancing resource efficiency and reducing latency. eccenca is part of the project amongst nineteen industry partners. Beyond building security and confidence for trustworthy edge and automated orchestration and maintenance, eccenca’s role in this project is to advance the graph generation and orchestration through AI techniques. The main goal is to create directed-acyclic property graphs of clusters and use them as data sources for AI/ML modules in cluster management. By doing so, the team is developing an intelligent management system that optimizes resource allocation, workloads, and data distribution across the edge-cloud continuum.

The Clever project is set to run until the conclusion of 2025. This ambitious initiative offers an additional year to enhance the platform and fully leverage its outcomes. Upon completion, all findings will be publicly disclosed. The project anticipates significant exploitable results that will distinguish it as a standout endeavor. This extended timeline facilitates further development, rigorous testing, and refinement of the platform, ensuring it achieves maximum impact and usability.

Innovating Distributed Knowledge Graphs

The emergence of the edge and heterogeneous compute continuum has highlighted the need for distributed, decentralized data solutions. A significant challenge is maintaining data consistency across distributed knowledge graphs. Traditional synchronization methods often lead to inconsistencies and require manual corrections.

The proposed solution of distributed Knowledge Graphs leverages immutable distributed ledger technology (DLT) to build consistent knowledge graphs across multiple nodes without centralized management. This method ensures synchronization of fact transactions through DLT's consensus algorithm, providing scalability, resilience, and consistency.

Advancing Distributed ML Model Registry and LLM Integration

Various artifact management tools were evaluated for use as the Distributed ML Model Registry (DMR), aiming to facilitate ML model management and deployment on the edge. Additionally, the team is assessing Large Language Models (LLMs) for knowledge retrieval tasks. This will enable interactive access to information stored on knowledge graphs through natural language, enhancing semantic lifting of specific data pipelines. The next step involves prototyping focusing on open-source availability, free licensing, ease of integration, authentication, and deployment simplicity.

Modeling and building Knowledge Graphs

The complexity of resource allocation and workload placement in distributed and heterogeneous cloud and edge environments requires innovative optimization tools. AI and deep learning (DL) solutions, though powerful, often act as black boxes. To address this, the team is using knowledge graphs to create practical, accurate, and explainable solutions. Knowledge Graphs offer a structured representation of knowledge, integrating entities, relationships, and attributes to model complex domains.

For CLEVER, the use of Knowledge Graph technology will aid resource allocation decisions, modeling both communication and computing aspects. By defining a data model that includes entities, relationships, and attributes, and storing in graph databases, we facilitate effective querying and graph mining functionalities, supporting knowledge discovery and decision-making.

We are excited to share these advancements and look forward to the transformative impact they will have on several European industries like smart agriculture, supply chain, augmented reality for shopping sites and the edge computing market.

eccenca has a scientific background and originated from the world-class research group of DBpedia. Ever since, eccenca has maintained close ties to researchers and research groups worldwide. The CLEVER Project is one of the four research projects, the eccenca team is currently contributing to. The CLEVER Project is driving innovation in edge computing, developing cutting-edge solutions to meet the needs of future intelligent systems. Stay tuned for more updates as we continue to push the boundaries of what is possible in edge-AI and intelligent computing. In the meanwhile, you can always check our project website:

Download the CEC Poster as PDF

Funded by the European Commission (Grant Agreement no. 101097560), CLEVER aims to revolutionize edge computing by enhancing resource efficiency, reducing latency, and positioning Europe as a leader in intelligent computing.

More news