Open Radio Access Networks (RAN) are revolutionizing the telecommunications industry by introducing disaggregated, programmable components and open interfaces. However, despite their promise, existing architectures lack robust support for real-time control, limiting their ability to address latency-sensitive operations effectively. To bridge this gap, we propose dApps—lightweight applications deployed directly on RAN nodes. Designed to extend the O-RAN architecture into real-time and user-plane domains, dApps overcome privacy and timing constraints by accessing data unavailable to traditional controllers. This enables them to execute control actions within sub-millisecond intervals, unlocking new possibilities for optimizing network slicing, scheduling, and resource management. To showcase the potential of dApps, we developed an open-source framework built on OpenAirInterface (OAI). This framework includes a reference architecture for dApps, covering their lifecycle from deployment to real-time interactions with RAN nodes. We also introduced the E3 interface, a communication protocol that integrates dApps with the broader O-RAN ecosystem, allowing seamless collaboration with other applications like xApps and rApps. Our framework has been validated in real-world 5G scenarios, demonstrating the feasibility of real-time control loops. Key use cases, such as spectrum sharing and user positioning, achieved control latencies below one millisecond, paving the way for intelligent, scalable, and efficient Open RAN systems.

A tutorial on how to setup the dApps with OpenAirInterface is available here.

If you use the dApp concept and/or the framework to develop your own dApps, please cite the following paper:

A. Lacava, L. Bonati, N. Mohamadi, R. Gangula, F. Kaltenberger, P. Johari, S. D’Oro, F. Cuomo, M. Polese, and T. Melodia, "dApps: Enabling Real-Time AI-Based Open RAN Control," arXiv:2501.16502 [cs.NI], pp. 1-31, January 2025. [pdf] [bibtex]