Cache Algorithm Optimization for SSD-Based Advertising Retrieval System

The paper presents SsdEngine, a hierarchical SSD‑based advertising retrieval cache that combines a thread‑local ART cache with a lock‑free central array, uses flying pointer‑swizzling to cut duplicate I/O, and evaluates eviction policies (LRU, DoubleClock, TinyLFU, ARC), showing LRU excels under Zipfian locality while TinyLFU improves small‑cache hit rates.

DoubleClockLRU evictionNVMe SSD

0 likes · 23 min read

Cache Algorithm Optimization for SSD-Based Advertising Retrieval System

UCloud Tech

Sep 2, 2019 · Backend Development

How UFS’s Performance‑Optimized Architecture Cuts I/O Latency to Sub‑10ms

This article explains how UCloud's performance‑oriented UFS leverages NVMe SSDs, upgrades to NFSv4, redesigns business indexing, and introduces a novel append‑only storage engine with stream/extent architecture to achieve sub‑10 ms latency and high IOPS for demanding AI and analytics workloads.

Distributed File SystemNFSv4NVMe SSD

0 likes · 17 min read

How UFS’s Performance‑Optimized Architecture Cuts I/O Latency to Sub‑10ms

Architects' Tech Alliance

Jun 26, 2017 · Databases

Huawei ES3600 v3 NVMe SSD Review and Performance Analysis

The review evaluates Huawei's ES3600 v3 enterprise NVMe SSD, detailing its specifications, benchmark methodology, and performance results across SQL Server, MySQL, and synthetic workloads, and concludes with an assessment of its strengths, weaknesses, and suitability for latency‑sensitive database applications.

Enterprise StorageHuaweiNVMe SSD

0 likes · 13 min read

Huawei ES3600 v3 NVMe SSD Review and Performance Analysis