The SMDB Workshop series sponsored by the IEEE TCDE Workgroup on Self-Managing Database Systems brings together researchers and practitioners to exchange ideas related to self-managing data management systems. After multiple years of successful SMDB workshops, we are organizing the SMDB Workshop once again, given the renewed interest in self-managing databases and distributed systems that focus on reducing the system management complexity. These systems are growing rapidly in scale and complexity, while skilled operational teams to run these systems have become rarer than ever. Furthermore, the agility with which enterprises want to move in the cloud can only be sustained via automation.
We are pleased to present a deep and diverse set of papers in this thirteenth edition of the SMDB Workshop. Four regular papers and two invited regular papers were submitted in total, which were subject to a rigorous review procedure; every paper had three to five independent reviews. This year we introduced a quick paper revision round that allowed the authors to incorporate the feedback from the reviewers into their papers resulting in high quality papers. Three papers were accepted without any revisions and three papers were accepted after the revision round for publication in this proceedings volume. The six regular papers were divided into two research sessions. We own our sincere gratitude to the members of the Technical Program Committee who contributed to assembling such a high-quality program, working within a very tight schedule.
The first paper in the first research session discusses several of the experiences and lessons learned while building Facebook’s smart warehouse systems by automating multiple optimizations, which were turned manually before. Particularly, these optimizations include knob tuning, materialized views, and partitioning. The second paper presents a solution for self-tuning databases using Bayesian-inference and proposes a set of improvements to the classic Bayesian approach. The third paper in this session proposes an adaptive methodology to handle updates with concurrent analytical queries (HTAP workload) given the known benefit of GPU processing for graph data analytics.
The first paper in the second research session explores trade-offs between ML models and streaming processing engines tuning, and studies the interaction of the configurations of the system knobs and the ML models. The second paper discusses the data placements policies in fog ecosystems to improve the latency, data freshness and data quality. The final paper in this session describes the design and implementation of AlphaSQL, an open-source, command-line tool to validate SQL queries and mitigate possible mistakes due to the complexity of the SQL schema, queries and the cyclic dependency among queries. The goal of AlphaSQL is to reduce the complexity and cost of SQL programming.
In addition to the papers, the workshop also features five keynotes talks. The first keynote talk is titled “Deep Data Integration” by Wang-Chiew Tan, Research Scientist, Facebook AI. The second keynote talk is titled “Towards instance-optimized data systems” by Tim Kraska, Associate Professor, MIT. The third talk is the Founder and Pioneers keynote talk by C. Mohan, Distinguished Visiting Professor, Tsinghua University, China with title “Modern Cloud DBMSs Vindicate Age-Old Work on Shared Disks DBMSs!”. The fourth keynote talk is titled “Accelerating Data Analytics in the Era of Ubiquitous Computing: Opportunities and Challenges” by Maya Gokhale, Distinguished Member of Technical Staff, Lawrence Livermore National Laboratory, USA. The final keynote is titled “Memory-Centric Computing” by Onur Mutlu, Professor of Computer Science, ETH Zurich. We thank and appreciate deeply all the invited keynote speakers. We would like to welcome all of you to attend SMDB 2022, which takes place online due to the directives involving COVID-19 and is part of the new resurgence that we are seeing across academia, research, and industry in developing self-managing systems.
Herodotos Herodotou
Yingjun Wu
SMDB 2022 Workshop General Chairs