Recently, the research team led by Professor Wang Xingwei from the School of Computer Science and Engineering achieved a significant breakthrough in large-scale model training and inference systems. Their latest research findings "ResCCL: Resource-Efficient Scheduling for Collective Communication" have been accepted and published at ACM SIGCOMM 2025. This marks the first time NEU has served as the first-affiliated institution for a paper presented at this premier international conference. Master student Liu Tongrui and doctoral student Hei Chenyang are the co-first authors of the paper, with Professors Li Fuliang and Wang Xingwei serving as co-corresponding authors.
From initial concept to production-level deployment, the team spent over a year of iterative refinement and rigorous experimental validation, ultimately achieving performance targets that far exceeded expectations. Based on this system, the team has initiated multiple follow-up research projects and leveraged its technical strengths to establish an in-depth research collaboration with Alibaba Cloud.
It is reported that ACM SIGCOMM (Special Interest Group on Data Communication) has been held for 39 consecutive years. As the most influential and longest-running international academic conference in the field of computer networking, it is hailed as the annual highlight of the computer networking community. The conference not only values the theoretical depth and practical foundation of research, but also emphasizes its tangible value for both academia and industry. Papers selected for inclusion serve as an epitome of the latest breakthroughs in computer networking and a barometer for future trends.
