Lune

SOSP2023Top-tier venue

Falcon: Fast OLTP Engine for Persistent Cache and Non-Volatile Memory

Zhicheng Ji, Kang Chen, Leping Wang, Mingxing Zhang, Yongwei Wu

2023Year
10Citations
3Top-tier citations

Abstract

Non-volatile memory(NVM) has the properties of both byte addressable and persistence, which provides new opportunities for building on-line transaction processing (OLTP) engines. Recently, a new feature called eADR puts CPU cache also in the persistence domain. Existing OLTP engines are based on volatile cache and now have the opportunity to improve performance further and reduce programming complexity with persistent cache.

This paper studies the impact of persistent cache on OLTP engines and revisits the existing designs. We have observed that naively removing the flush instructions can trigger the write amplification because of the granularity mismatch between the cache line and NVM access. We propose Falcon, a new OLTP engine for eADR-enabled NVM. Falcon is based on the in-place update architecture. The small log window design in Falcon avoids the NVM writes while logging. The selective data flush design reduces the data flush and the write amplification while flushing data. Evaluations show that under TPC-C workloads, Falcon achieves 1.21× ∼ 1.35× improvement over the state-of-the-art OLTP engine.

Ask about this paper

Your agent reads all of it.

Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.

Questions to start from

Your agent calls

Luneget_paper_fulltext

Ask in Lune

Free to start. No credit card required.

lune papers fulltext 97aa3cd9-95b2-4bbf-83d5-711c881d3b97

Cited by top-tier papers3

Ask how each one uses it

Builds on17

Related papers

Dusk over the sea between two cliffs drawn in fine vertical lines