Researchers from Smartness took part in the 38th edition of the IEEE/IFIP Network Operations and Management Symposium (NOMS 2025), held from May 12 to 16 in Honolulu, Hawaii. During the event, they shared research exploring how applications and services can be better executed in different environments, including edge computing and mobile devices.
Prof. Fábio L. Verdi (UFSCar) presented two studies as part of the Research Strand DisCoNet. The first, titled “DECOMPOSER: Functional decomposition and distributed execution of monolithic applications to heterogeneous resources,” by Guilherme Matos, Washington Dias, Andrew Williams, Fábio L. Verdi, was accepted as a full paper in the Technical Sessions. This study explores how large, monolithic applications can be broken down into modular components to run on various hardware platforms. The approach leverages LLVM and MLIR to identify functions for distribution, achieving improved memory use and reduced cache pressure in both data center and edge computing scenarios.
The second work, showcased in the Demos session, was “Demonstrating the Advantages of Computational Offloading of XR Services via WebAssembly,” by Gabriel Espindola, Matheus Lucas, Gustavo Spadotto, Cristiano Both, Bruno Silvestre, Kleber Cardoso, Andrew Williams, Sand L. Correa, and Fábio L. Verdi. This work highlights how XR (extended reality) services can be offloaded to edge computing environments using WebAssembly. The demonstration showed improved performance and reduced heat generation on mobile devices when running object detection tasks.
Additionally, Prof. Raza Ul Mustafa (Loyola University New Orleans) presented the work “Investigating the Impact of Channel Metrics in 5G NSA and SA on Video Streaming QoE,” by Raza Ul Mustafa, Md Tariqul Islam, and Christian Esteve Rothenberg at the Mini-Conference. This study, part of the Research Strand C3LA, examined how channel-level metrics such as RSRP, RSRQ, and SNR in 5G networks influence the quality of video streaming, comparing the performance of 5G Standalone and Non-Standalone deployments in different indoor and outdoor settings.
If you’d like to dive deeper or explore these studies further, the papers presented at the event are openly available on the Smartness Zenodo community:
You can also visit the Smartness Zenodo Community to discover other studies and stay updated on the latest research.