Bai, ShuhanShuhanBaiWan, HuHuWanHuang, YunYunHuangSun, XuanXuanSunWu, FeiFeiWuXie, ChangshengChangshengXieHsieh, Hung ChihHung ChihHsiehTEI-WEI KUOXue, Chun JasonChun JasonXue2023-08-072023-08-072023-01-0102780070https://scholars.lib.ntu.edu.tw/handle/123456789/634446Big data applications, such as recommendation system and social network, often generate a huge number of fine-grained reads to the storage. Block-oriented storage devices upon the traditional storage system rely on the paging mechanism to migrate pages to the host DRAM, tending to suffer from these fine-grained read operations in terms of I/O traffic as well as performance. Motivated by this challenge, an efficient fine-grained read framework, Pipette, is proposed in this paper as an extension to the traditional I/O framework. With adaptive design for caching, merging and scheduling, Pipette explores locality and acceleration for fine-grained read requests to establish an efficient byte-granular read path upon the dedicated byte-addressable interface. When the Pipette prototype on an SSD runs popular workloads, we measured throughput gains by up to 50% and 54% with traffic reduction in the range of 41.3× and 56.5×.file system | fine-grained reads | Memory management | Merging | Parallel processing | Performance evaluation | Random access memory | Recommender systems | solid-state drive | Throughput[SDGs]SDG11journal article10.1109/TCAD.2023.32765202-s2.0-85162904924https://api.elsevier.com/content/abstract/scopus_id/85162904924