What is it about?
This is a tutorial where we present recent advances in developing database systems for resource disaggregation. Specifically, we cover two forms of disaggregation---that of storage, which is widely adopted in today's cloud-native database systems, and that of memory, which is an emerging trend in cloud data centers, and discuss the enabling hardware and networking techniques (e.g., RDMA, persistent memory, and CXL) and database system designs (e.g., buffer management, indexing, and fault tolerance) for the novel cloud architectures.
Featured Image
Why is it important?
Recent years have witnessed a wave of cloud-native database systems designed for storage disaggregation, and a line of more recent work is pushing the design toward memory disaggregation. It's time for database researchers and practitioners to recognize the trends, opportunities, and challenges. However, there is a lack of systematic investigation into state-of-the-art work on disaggregated database systems.
Perspectives
To disaggregate, or not to disaggregate? We should now ask this question when building cloud-scale hardware and software systems. This tutorial seeks to raise the awareness and provide a fairly comprehensive view of recent work on database systems for this topic.
Qizhen Zhang
University of Toronto
Read the Original
This page is a summary of: Disaggregated Database Systems, June 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3555041.3589403.
You can read the full text:
Contributors
The following have contributed to this page
