Faculty Knowledge Exchange (KE) Award 2023

Cloud Computing has become an indispensable infrastructure for the digital economy and a symbol of competitiveness for any economic entity. Optimisation plays an irreplaceable role in advancing cloud computing technologies for promising and rapidly growing cloud computing businesses such as live streaming. As of June 2021, the number of online live-streaming users in China reached 638 million, and it is projected that the global market size of this industry will climb from USD 59.14 billion in 2021 to more than USD 330 billion by 2030. To minimise operational cost of live streaming, reducing the bandwidth cost charged by Internet Service Providers is crucial.

To address this fundamental issue, Professor YUAN and his team, including PhD students Mr Pengxiang ZHAO and Mr Hanyu HU from HKU, in collaboration with Huawei Cloud, have developed an intelligent bandwidth allocation system called GSCO. Equipped with various cutting-edge optimisation algorithms developed by Professor Yuan, GSCO has been successfully deployed by Huawei Cloud for its daily B2B live streaming business serving for over 15 million customers in more than 60 countries since 2020, with a remarkable reduction of approximately 30% in total bandwidth cost (equivalent to over USD 49.6 million), a cut of 98.6% manpower, and a ten-fold increase of live streaming business volume.

The impact of Professor Yuan's work on advancing knowledge and fostering innovation in the industry has been widely recognised, and his team has recently been awarded as one of the six finalists for the prestigious Franz Edelman Award in 2023. This achievement is particularly noteworthy as his team is the first-ever finalist from the Cloud Computing area in the past 52 years, garnering significant media attention including People’s Daily and Sina.com.

Details of the Journal papers:

1. Yuan, X., Zhao, P., Hu, H., You, J., Yang C., Peng W., Kang Y., and Teo K., Huawei Cloud Adopts Operations Research for Live Streaming Services to Save Network Bandwidth Cost: The GSCO System, Informs Journal of Applied Analytics, to appear.
2. He, B., Ma, F., Xu, S., and Yuan, X. (2022). A generalized primal-dual algorithm with improved convergence condition for saddle point problems. SIAM Journal on Imaging Sciences, 15(3), 1157-1183.
3. He, B., and Yuan, X. (2021). Balanced augmented Lagrangian method for convex programming. arXiv preprint arXiv:2108.08554.
4. He, B., and Yuan, X. (2018). A class of ADMM-based algorithms for three-block separable convex programming. Computational Optimization and Applications, 70, 791-826.