Front Matter

Lei Chen, Fatma ÖzcanPVLDB 2019

This document serves as the editorial introduction and organizational roadmap for Volume 12, Issue 7 of the Proceedings of the VLDB Endowment (PVLDB), published in March 2019. It highlights eight peer-reviewed articles presented at the 45th International Conference on Very Large Data Bases (VLDB 2019) in Los Angeles.

Proceedings of the VLDB Endowment

The PVLDB Publication Model

The "Front Matter" outlines the unique publication model of PVLDB, which addresses the limitations of traditional academic conference cycles. Conventional models often suffer from long lead times and binary "accept/reject" decisions. PVLDB employs a hybrid approach:

  • Continuous Submission: Twelve monthly deadlines per year.
  • Agile Reviewing: A team of 166 expert researchers and 17 Associate Editors ensures a rapid four-week review turnaround.
  • Revision Cycle: Unlike many conferences, this model includes a revision phase, allowing authors to improve paper quality based on feedback before final publication.

This "journal-style" model combines the rigor of a journal with the visibility and community engagement of a major annual conference.

The Program Chairs, Lei Chen and Fatma Özcan, identify several key trends within this issue that reflect the evolving needs of the data management field:

  • Convergence of Machine Learning and Databases: There is a significant shift toward integrating ML directly into database engines, including declarative recursion in RDBMS and hardware acceleration for generalized linear models.
  • Elasticity in Stream Processing: As data volumes fluctuate, "elastic" distributed stream processing—systems that scale dynamically—has become a core research priority.
  • Data Quality in Complex Structures: Research is moving beyond simple table cleaning to address "fixing" complex graph structures, driven by the rise of graph databases in social and biological sciences.
  • Mobile-First Recovery: With the ubiquity of mobile devices, techniques like "database shadowing" are emerging as critical for crash recovery in mobile DBMS.
VLDB 2019 Conference

This issue presents eight distinct research papers (pages 724–835) covering a variety of technical challenges:

  • Cost Optimization: Strategies for reducing scaling operations in distributed stream processing to minimize cloud infrastructure costs.
  • Data Provenance: Introduction of "ProvCite," a system designed for provenance-based data citation.
  • Clustering at Scale: New algorithms for k-center clustering with outliers that maintain high accuracy in MapReduce and streaming environments.
  • Performance Speedups: "MLWeaving," a system for "any-precision learning" that leverages hardware acceleration to optimize generalized linear models.

Conclusion

The editorial letter underscores the success of the PVLDB model in fostering high-quality, high-velocity research. While maintaining such a process requires significant effort from the review board, it serves as a vital template for other fast-moving computer science sub-fields seeking to balance rigor with rapid innovation.

Summary updated

BibTeX
@article{chen2019front,
  author = {Lei Chen and Fatma Özcan},
  title = {{Front Matter}},
  journal = {Proc. VLDB Endow.},
  year = {2019},
}
Figure