Jiaxiao Zheng received the Best Student Paper Award at the 17th International Symposium on Modeling and Optimization in Mobile, Ad Hoc and Wireless Networks (WiOpt 2019). The conference took place this summer in Avignon, France.
WiOpt solicits “high-quality contributions [from] a variety of perspectives," including: performance analysis, protocol design, wireless communication, and optimization theory. The symposium looks for papers that “improve the state-of-the-art in the analysis, dimensioning and operations of wireless networks by providing insights into theoretical aspects as well as providing practical methods and tools.”
Zheng co-authored the winning paper, “Elastic multi-resource network slicing: Can protection lead to improved performance,” with his advisor, Prof. Gustavo de Veciana. The paper explored resource provisioning in the context of network slicing in a multi-tenant mobile edge computing network.
Inspired by classical alpha-fairness, the researchers developed a sharing criterion—alpha-SCS (Share Constrained Slicing), where each slice of the network gets a predefined network share, and re-distributes its share across its users. By tuning parameter values, alpha-SCS can achieve different trade-offs among intra- and inter-slice fairness and overall efficiency. The researchers also showed that alpha-SCS achieves stability under fairly weak conditions in a setting with stochastic load and elastic user service requirements. Extensive simulations demonstrated that alpha-SCS achieves inter-slice protection as well as improving job delay and user-perceived throughput.
As a WNCG student, Zheng researched sharing, allocation, and provisioning in mobile network slicing.
His main interest is in devising low-complexity resource sharing framework across tenants to accommodate their highly dynamic user demands. “It's important, as well as challenging,” he remarked.
Zheng graduated recently, obtaining his Ph.D. this past spring before joining Google as a software engineer. Based in Seattle, he now works on an open-sourced project for Google’s Cloud AI platform geared at leveraging the power of ML/AI through cloud computing.