publications

2026

1. NSDI

   Libra: Flexible Request Partitioning and Scheduling for Serving Unbalanced and Dynamic LLM Workloads

   Chaoyi Ruan, Yinhe Chen, Dongqi Tian, and 4 more authors

   In Proc. of the USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2026
2. NSDI

   Cortex: Achieving Low-Latency, Cost-Efficient Remote Data Access For LLM via Semantic-Aware Knowledge Caching

   Chaoyi Ruan, Chao Bi, Kaiwen Zheng, and 3 more authors

   In Proc. of the USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2026
3. ICLR

   Revisiting Parameter Server in LLM Post-Training

   Xinyi Wan, Penghui Qi, Guangxing Huang, and 3 more authors

   In Proc. of the International Conference on Learning Representations (ICLR), 2026

2025

1. VLDB

   Adapting to Data Affinity Changes in Geo-Replicated Database via Row-Level Paxos- Group Affiliation Re-Assignment

   Chaoyi Ruan, Yingqiang Zhang, Juncheng Zhang, and 6 more authors

   Proc. VLDB Endow., 2025
2. ICCD

   DHeLlam: General-Purpose, Automatic Micro-batch Co-execution for Distributed LLM Training

   Haiquan Wang^*, Chaoyi Ruan^*, Jia He, and 4 more authors

   In Proc. of the IEEE International Conference on Computer Design (ICCD), 2025

   Awarded

   Best Paper Award

2023

1. ASPLOS

   Persistent Memory Disaggregation for Cloud-Native Relational Databases

   Chaoyi Ruan, Yingqiang Zhang, Chao Bi, and 7 more authors

   In Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 3, Vancouver, BC, Canada, 2023

   DOI

2021

1. VLDB

   Towards cost-effective and elastic cloud database deployment via memory disaggregation

   Yingqiang Zhang^*, Chaoyi Ruan^*, Cheng Li, and 8 more authors

   Proc. VLDB Endow., Jun 2021

   DOI
2. FAST

   SpanDB: A fast, Cost-Effective LSM-tree based KV store on hybrid storage

   Hao Chen, Chaoyi Ruan, Cheng Li, and 2 more authors

   In 19th USENIX Conference on File and Storage Technologies (FAST 21), Jun 2021
3. TOS

   Leveraging NVMe SSDs for Building a Fast, Cost-effective, LSM-tree-based KV Store

   Cheng Li, Hao Chen, Chaoyi Ruan, and 2 more authors

   ACM Trans. Storage, Oct 2021

   Abs DOI

   Key-value (KV) stores support many crucial applications and services. They perform fast in-memory processing but are still often limited by I/O performance. The recent emergence of high-speed commodity non-volatile memory express solid-state drives (NVMe SSDs) has propelled new KV system designs that take advantage of their ultra-low latency and high bandwidth. Meanwhile, to switch to entirely new data layouts and scale up entire databases to high-end SSDs requires considerable investment. As a compromise, we propose SpanDB, an LSM-tree-based KV store that adapts the popular RocksDB system to utilize selective deployment of high-speed SSDs. SpanDB allows users to host the bulk of their data on cheaper and larger SSDs (and even hard disc drives with certain workloads), while relocating write-ahead logs (WAL) and the top levels of the LSM-tree to a much smaller and faster NVMe SSD. To better utilize this fast disk, SpanDB provides high-speed, parallel WAL writes via SPDK, and enables asynchronous request processing to mitigate inter-thread synchronization overhead and work efficiently with polling-based I/O. To ease the live data migration between fast and slow disks, we introduce TopFS, a stripped-down file system providing familiar file interface wrappers on top of SPDK I/O. Our evaluation shows that SpanDB simultaneously improves RocksDB’s throughput by up to 8.8times and reduces its latency by 9.5–58.3%. Compared with KVell, a system designed for high-end SSDs, SpanDB achieves 96–140% of its throughput, with a 2.3–21.6times lower latency, at a cheaper storage configuration.

4. EuroSys

   Achieving low tail-latency and high scalability for serializable transactions in edge computing

   Xusheng Chen, Haoze Song, Jianyu Jiang, and 6 more authors

   In Proceedings of the Sixteenth European Conference on Computer Systems, Online Event, United Kingdom, Oct 2021

   Abs DOI

   A distributed database utilizing the wide-spread edge computing servers to provide low-latency data access with the serializability guarantee is highly desirable for emerging edge computing applications. In an edge database, nodes are divided into regions, and a transaction can be categorized as intra-region (IRT) or cross-region (CRT) based on whether it accesses data in different regions. In addition to serializability, we insist that a practical edge database should provide low tail latency for both IRTs and CRTs, and such low latency must be scalable to a large number of regions. Unfortunately, none of existing geo-replicated serializable databases or edge databases can meet such requirements.In this paper, we present Dast (Decentralized Anticipate and STretch), the first edge database that can meet the stringent performance requirements with serializability. Our key idea is to order transactions by anticipating when they are ready to execute: Dast binds an IRT to the latest timestamp and binds a CRT to a future timestamp to avoid the coordination of CRTs blocking IRTs. Dast also carries a new stretchable clock abstraction to tolerate inaccurate anticipations and to handle cross-region data reads. Our evaluation shows that, compared to three relevant serializable databases, Dast’s 99-percentile latency was 87.9% 93.2% lower for IRTs and 27.7% 70.4% lower for CRTs; Dast’s low latency is scalable to a large number of regions.

5. PerfEstimator: A Generic and Extensible Performance Estimator for Data Parallel DNN Training

   Chengru Yang, Zhehao Li, Chaoyi Ruan, and 4 more authors

   In 2021 IEEE/ACM International Workshop on Cloud Intelligence (CloudIntelligence), Oct 2021

   DOI

2018

1. infocomm

   AppDNA: App Behavior Profiling via Graph-based Deep Learning

   Shuangshuang Xue, Lan Zhang, Anran Li, and 3 more authors

   In IEEE INFOCOM 2018 - IEEE Conference on Computer Communications, Oct 2018

   DOI