LF Networking has announced Essedum Release 1.0, a modular open source platform designed to accelerate the integration of AI into network applications. The platform supports data connections, pipeline management, and model implementation for both on-premise and cloud environments.
Essedum 1.0 includes several core features. Connections provide communication links between software systems to enable data exchange. Datasets support the import and management of data from various sources, including storage buckets, MySQL databases, and REST APIs.
Pipelines enable the building and management of both training and inferencing workflows for AI/ML workloads. This includes model optimization and deployment. The Models feature provides access to AI models from configured connections across different platforms.
Modular platform for AI networking
The Essedum project was introduced this year by LFN member Infosys. The platform offers a comprehensive framework that covers the entire chain: from data ingestion to pipeline orchestration and model deployment. This provides developers and operators with tools to build AI-driven network solutions efficiently.
The first release introduces platform capabilities for secure data connections, pipeline creation, model management, and multi-platform deployment. The system supports various environments, from on-premise servers to cloud services such as AWS SageMaker, Azure ML, and GCP Vertex AI.
In addition, the system offers Endpoints for viewing and managing all connected endpoints from a single central interface. Adapters simplify integration with external services without the need to configure host details. The Remote Executor can run pipelines or programs on external servers for compute-intensive processing.
Future plans
Several improvements are planned for upcoming releases. Docker and Helm-based deployment automation should facilitate implementation. Support for reading PDF and Excel files will also be added.
Secrets management and extensive role-based access control are also on the roadmap. In addition, support for public cloud platforms will be further expanded.
In addition to the new release, the Essedum community has built a Sandbox instance in collaboration with the University of New Hampshire. This test environment is available to anyone who wants to duplicate the environment to try out Essedum.
Tip: Europe’s open source strategy falters despite widespread adoption