Linux Foundation Launches EdgeX Foundry: An Open Framework for Edge Computing and IoT

The Linux Foundation has launched EdgeX Foundry, a project aimed at establishing an open framework for IoT computing and an ecosystem of interoperable components.

EdgeX Foundry aims to promote an edge‑computing model where enabled devices do not send their data back to the cloud for processing, but instead rely on local gateway devices for computation and analysis. This approach’s main advantages are reduced latency and network traffic, and increased security. It is considered especially valuable for industrial applications such as power generation and smart traffic lights that do not require cloud components, as well as more general use cases where data from multiple devices is collected locally before being sent to the cloud.

According to the announcement, Dell will contribute the source code of its Fuse framework under the Apache 2.0 license; the framework consists of more than ten micro‑services and over 125,000 lines of code. The project will also receive code contributions from other members, including AMD, Ubuntu, Canonical, VMWare and about 50 other companies, although details are still unclear.

InfoQ has spoken with Philip DesAutels, senior IoT director at the Linux Foundation.

When will developers be able to use the source code? Which parts of the framework will be available first, and what comes next?

DesAutels: The code will be released as EdgeX on or before May 31. Because we are converting 125 K lines of code, we are undertaking a large‑scale transformation and a namespace renaming exercise from FUSE to EdgeX. Companies that need earlier access can follow a simple process, and dozens have already done so. In short, the code is being transformed and changes are happening.

Can you give a high‑level overview of the capabilities the framework will provide to developers?

DesAutels: The best way is to look at our architecture slides:

We provide a unified framework through four core APIs (Security, Management, Device, and Service). The ecosystem of companies that participate and support EdgeX will contribute real services that can be assembled into EdgeX Platforms (implementations). In this sense, EdgeX will offer example device services, support services, management tools, etc., that can be used as reference implementations. We also expect some of these components to mature to production‑ready levels over time as part of the open‑source EdgeX Foundry project.

Will EdgeX provide a single, cohesive API or will it be more of a loosely coupled set of components?

DesAutels: EdgeX provides a core bus that unifies the four APIs into one framework, allowing loosely coupled components to interact in a coherent and consistent way.

What does a typical architecture of an EdgeX‑driven solution look like?

DesAutels: In the simplest implementation, a single gateway device runs the EdgeX core along with a set of device, support, and export services as well as management and security services. This becomes an edge processing node that interacts with multiple devices, sensors, machines, etc., before connecting to the cloud.

What domains do you especially envision solutions based on EdgeX for? What typical requirements suit EdgeX?

DesAutels: The community is building small‑footprint installation scripts, and the framework’s flexibility and a growing ecosystem mean we should expect a wide range of applicability—from industrial to civil infrastructure, enterprises, and even consumer use cases.

The Linux Foundation’s plan is not the first attempt to create a standard platform for interconnecting IoT devices. Previous examples include Cisco’s iOx platform; Pacific Northwest National Laboratory’s Android‑based reference platform Kaval; and the European Telecommunications Standards Institute’s MEC specification. Whether EdgeX Foundry will achieve greater success remains to be seen.