Allbesmart: Democratizing AI-RAN Experimentation with Open Source

Allbesmart, Lda,  is an innovation-driven company dedicated to advancing wireless communications and smart technologies through cutting-edge research and development. Paulo Marques, founder of Allbesmart, Lda, holds a doctorate in wireless communications and is a professor at the Polytechnic University of Castelo Branco in Portugal. He also helms business development at Allbesmart, focusing on transforming the company’s R&D efforts into innovative products. In a recent exchange about AI-RAN Alliance membership, we asked Paulo to introduce us all to Allbesmart, Lda.

Meet Allbesmart, Lda

Established in 2015, Allbesmart, Lda, is an engineering company specializing in open-source software-defined 3GPP implementations. The company is an active contributor to the open-source OpenAirInterface™ (OAI) protocol stack. Allbesmart develops and maintains the OAIBOX™ platform, an abstraction layer to facilitate the use and deployment of OAI, widely used for training, research, and validation of next-generation wireless technologies. The OAIBOX™ platform, integrated with NVIDIA GPUs, enables AI models to be deployed at the edge, for end-to-end testing, experimentation, and benchmarking against conventional approaches.

Allbesmart’s team has collaborated for several years with NVIDIA and the OAI Software Alliance on integrating NVIDIA Aerial L1 with the OAI L2/L3 stack.

A Shared Mission of Transformation

The mission of the Alliance is to transform the telecom industry by integrating AI with RAN for innovation, increased efficiency, and new economic opportunities. Allbesmart’s mission is to democratize wireless experimentation by promoting the adoption of open-source technologies and open standards for 5G and beyond. In this context, we aim to enable end-to-end experimentation and validation of AI-RAN research ideas and use cases,” explains Paulo Marques. “I see AI as essential for making RANs more sustainable and efficient. It can reduce costs, optimize energy and spectrum use, and enable edge applications that create new revenue opportunities such as integrated sensing and communication.”

Allbesmart operates in the test and measurement market and is interested in AI-RAN Alliance’s Test Methodology flagship initiative—particularly in exploring how the open-source OAIBOX™ platform can contribute to advancing experimentation and benchmarking in the AI-RAN domain.

Eyes on the Future

Paulo Marques shared that Allbesmart is investing in R&D in two areas of opportunity:

  • The shared use of GPU infrastructure between the RAN and AI processes not related to the RAN, as shown in this video with the NVIDIA Aerial 
  • The use of integrated sensing and communication (ISAC) for surveillance applications on the network edge, enabling “see-through-wall” capabilities, as demonstrated in this video 

Learn More About AI-RAN Alliance Membership

The AI-RAN Alliance thrives on collaboration across the telecom and AI communities, and these communities thrive on the collaboration fostered by the Alliance. To learn how your organization can influence the future of AI-RAN, visit our membership page.