Xin Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xin Li has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Jiawei Huang
Central South University
2 shared publications Changsheng Zhong
1 shared publication- Jiacheng Xie
Central South University, Changsha, China
1 shared publication - Jianxin Wang
School of Computer Science and Engineering, Central South University, Changsha, Hunan, China
1 shared publication Jiawei Huang
1 shared publication- Jiaxin Jiang1 shared publication
Latest Publications
- Open Access
- 27 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 1019-1029
PipeTree: Decision Tree Partitioning for Efficient Inference on Programmable Switches
- 27 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 1019-1029
Open AccessPipeTree: Decision Tree Partitioning for Efficient Inference on Programmable Switches
- Open Access
- 15 August 2026
- Published by Elsevier in SSRN Electronic Journal
Achieving Predictable Performance for Long-Haul RDMA using Token Loop between Data Centers
Remote Direct Memory Access (RDMA) is the most promising transport technology for high-performance cross-datacenter applications in modern cloud infrastructure. However, deploying RDMA over long-haul links introduces a huge Bandwidth-Delay Product (BDP), much larger than the buffer capacity of current datacenter switch. Though current congestion control mechanisms commonly provide low-latency transmission within a single datacenter, they cannot effectively handle long-haul scenarios, which suffer from slow feedback loops and frequent switch buffer overflow. In this paper, we propose Ferryman, a systematic solution for long-haul RDMA based on a token-loop mechanism. Our key insight is that the available bandwidth should be precisely predicted in advance to avoid blind bandwidth exploration. To this end, we maintain a token loop across inter-DC links and incorporate a bandwidth allocation algorithm to rapidly calibrate sender rates to available bandwidth. We conduct extensive experiments on both testbed and NS3 simulations. The evaluation results show that, compared with state-of-the-art designs, Ferryman significantly mitigates cross-DC congestion and improves overall throughput while effectively preventing switch buffer overflow.
- 15 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessAchieving Predictable Performance for Long-Haul RDMA using Token Loop between Data Centers
Remote Direct Memory Access (RDMA) is the most promising transport technology for high-performance cross-datacenter applications in modern cloud infrastructure. However, deploying RDMA over long-haul links introduces a huge Bandwidth-Delay Product (BDP), much larger than the buffer capacity of current datacenter switch. Though current congestion control mechanisms commonly provide low-latency transmission within a single datacenter, they cannot effectively handle long-haul scenarios, which suffer from slow feedback loops and frequent switch buffer overflow. In this paper, we propose Ferryman, a systematic solution for long-haul RDMA based on a token-loop mechanism. Our key insight is that the available bandwidth should be precisely predicted in advance to avoid blind bandwidth exploration. To this end, we maintain a token loop across inter-DC links and incorporate a bandwidth allocation algorithm to rapidly calibrate sender rates to available bandwidth. We conduct extensive experiments on both testbed and NS3 simulations. The evaluation results show that, compared with state-of-the-art designs, Ferryman significantly mitigates cross-DC congestion and improves overall throughput while effectively preventing switch buffer overflow.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 6944-6955
Cardinality is Not Enough: Super Host Detection via Segmented Cardinality Estimation
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 6944-6955
Open AccessCardinality is Not Enough: Super Host Detection via Segmented Cardinality Estimation