So Far and yet so Near - Accelerating Distributed Joins with CXL
Open Access
- 9 June 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-9
Abstract
Distributed partitioned joins are one of the most expensive operators in distributed DBMSs where a major part of the execution is attributed to network transfer costs. Although high-speed network technologies, such as RDMA, can lower this cost, they still come with significantly higher latency than local DRAM access. The emerging CXL interconnect protocol promises to provide direct and cache-coherent access to remote memory while offering byte-addressable memory access without CPU intervention. For short-distance communication in distributed DBMSs, CXL represents an interesting alternative for low-latency requirements. In this work, we explore how CXL can be leveraged for engine-internal communication and data exchange. We discuss and apply communication strategies to distributed joins. We emulate various CXL characteristics based on optimistic and pessimistic assumptions on the real performance of upcoming CXL devices and evaluate their impact on the execution of distributed joins. Our results show that CXL has the potential to improve distributed join performance.Keywords
Funding Information
- Actian Corp. (Exploiting Persistent Memory)
This publication has 25 references indexed in Scilit:
- CXL over Ethernet: A Novel FPGA-based Memory Disaggregation Design in Data Centers2023
- Computational CXL-Memory Solution for Accelerating Memory-Intensive ApplicationsIEEE Computer Architecture Letters, 2022
- Scaling of Memory Performance and Capacity with CXL Memory Expander2022
- Thinking More about RDMA Memory Semantics2021
- Low-Latency Communication for Fast DBMS Using RDMA and Shared Memory2020
- Joins in a heterogeneous memory hierarchy2018
- VectorH2016
- Bitwise dimensional co-clustering for analytical workloadsThe VLDB Journal, 2016
- Vectorization vs. compilation in query execution2011
- Hash table collision resolution with direct chainingJournal of Algorithms, 1989