Tianming Yang
6 Papers
Tianming Yang is an academic researcher. The author has contributed to research in topics: Cache & Computer science. The author has an hindex of 4, co-authored 6 publications.
Chat about Author
Papers
Lifelong Disk Failure Prediction via GAN-Based Anomaly Detection
Tianming Jiang,Jiangfeng Zeng,Ke Zhou,Ping Huang,Tianming Yang +4 more
- 01 Nov 2019
TL;DR: Experimental results on real-world SMART datasets demonstrate that, compared with the state-of-the-art supervised machine learning based methods, this approach predicts disk failures at a higher accuracy for the entire lifetime of models, i.e., both the initial period and the long-term usage.
31
Demystifying Cache Policies for Photo Stores at Scale: A Tencent Case Study
Ke Zhou,Si Sun,Hua Wang,Ping Huang,Xubin He,Rui Lan,Wenyan Li,Wenjie Liu,Tianming Yang +8 more
- 12 Jun 2018
TL;DR: This work presents a comprehensive study on internet-scale photo caching algorithms in the case of QQPhoto from Tencent Inc., the largest social network service company in China, and proposes to incorporate a prefetcher in the cache stack based on the observed immediacy feature that is unique to the QQ photo workload.
23
CARS: A Multi-layer Conflict-Aware Request Scheduler for NVMe SSDs
Tianming Yang,Ping Huang,Weiying Zhang,Haitao Wu,Longxin Lin +4 more
- 25 Mar 2019
TL;DR: This work proposes a Conflict Aware Request Scheduling policy named CARS for NVMe SSDs to maximally leverage the rich parallelism available in modern NV me SSDs, and shows that the scheduler can reduce the slowdown metric by up to 46% relative to the de facto round-robin scheduling policy.
15
Improving Cache Performance for Large-Scale Photo Stores via Heuristic Prefetching Scheme
TL;DR: This work presents a comprehensive study on internet-scale photo caching algorithms in the case of QQPhoto from Tencent Inc., the largest social network service company in China, and proposes to incorporate a prefetcher in the cache stack based on the observed immediacy feature that is unique to the QQ photo workload.
12
A Shingle-Aware Persistent Cache Management Scheme for DM-SMR Disks
Tianming Yang,Haitao Wu,Ping Huang,Fei Zhang +3 more
- 01 Nov 2017
TL;DR: Leveraging the sequential-only write property of shingled devices, this paper proposes a new shingle aware persistent cache cleaning policy for DM-SMR drives, which delivers better performance (by up to 2.5X) via dramatically reducing write amplification associated with persistence cleaning and alleviating fragmented reads.
7