28. Mai – 02. Juni 2017, Dagstuhl-Seminar 17222

Robust Performance in Database Query Processing


Renata Borovica-Gajic (The University of Melbourne, AU)
Goetz Graefe (Google – Madison, US)
Allison Lee (Snowflake Computing Inc. – San Mateo, US)

Auskunft zu diesem Dagstuhl-Seminar erteilt

Dagstuhl Service Team


Dagstuhl Report, Volume 7, Issue 5 Dagstuhl Report
Dagstuhl's Impact: Dokumente verfügbar


The Dagstuhl Seminar 17222 on "Robust performance in database query processing" assembled researchers from industry and academia for the third time to discuss robustness issues in database query performance. The seminar gathered 24 researchers around the world working on plan generation and plan execution in database query processing and in cloud-based massively parallel systems with the purpose to address the open research challenges with respect to the robustness of database management systems.

Delivering robust query performance is well known to be a difficult problem for database management systems. All experienced DBAs and database users are familiar with sudden disruptions in data centers due to poor performance of queries that have performed perfectly well in the past. The goal of the seminar is to discuss the current state-of-the-art, to identify specific research opportunities in order to improve the state-of-affairs in query processing, and to develop new approaches or even solutions for these opportunities.

Unlike the previous seminars, the organizers (Renata Borovica-Gajic, Goetz Graefe and Allison Lee) this time attempted to have a focused subset of topics that the participants discussed and analyzed in more depth. From the proposed topics on algorithm choices, join sequences, updates, database utilities, parallelism and skew, column stores, physical database design, and explainability of non-robust query performance, the participants chose four topics and formed four work groups: i) one discussing updates and database utilities, ii) one discussing parallelism and skew, iii) one discussing join sequences, and iv) one focusing on the explanations of the sources of non-robust performance.

Upon choosing the topics of interest, the organizers then guided the participants to approach the topic through a set of steps: by first considering related work in the area; then introducing metrics and tests that will be used for testing the validity and robustness of the solution; after metrics, the focus was on proposing specific mechanisms for the proposed approaches; and finally the last step focused on the implementation policies.

The seminar thus spent its first day on reviewing prior related work, with a special emphasis on the pieces of work that appeared following the previous instances of the seminar: benchmarks (Dagstuhl 12321), Smooth Scan , and Generalized join.

Tuesday was spent on defining metrics and tests. On Wednesday, the participants discussed possible alternative approaches and hiked together in the woods. Thursday was focused on driving one chosen approach to specific mechanisms. Finally, we spent Friday on discussing the policies and presented the overall progress.

At the end of the week, each group was hoping to continue their work towards a research publication. The group on parallelism and skew was hoping to publish first a survey on forms of skew and existing remedies for skew. The work group on dynamic join sequences even had a working prototype by the end of the seminar. The reports of work groups are presented next.

Summary text license
  Creative Commons BY 3.0 Unported license
  Renata Borovica-Gajic, Goetz Graefe, Allison Lee, and Glenn Paulley

Dagstuhl-Seminar Series


  • Data Bases / Information Retrieval


  • Database
  • Query processing
  • Query optimization
  • Query execution
  • Map-reduce
  • Systems
  • Cloud computing
  • Performance
  • Scalability
  • Robustness
  • Reliability
  • Predictability
  • Planning
  • Uncertainty


In der Reihe Dagstuhl Reports werden alle Dagstuhl-Seminare und Dagstuhl-Perspektiven-Workshops dokumentiert. Die Organisatoren stellen zusammen mit dem Collector des Seminars einen Bericht zusammen, der die Beiträge der Autoren zusammenfasst und um eine Zusammenfassung ergänzt.


Download Übersichtsflyer (PDF).

Dagstuhl's Impact

Bitte informieren Sie uns, wenn eine Veröffentlichung ausgehend von Ihrem Seminar entsteht. Derartige Veröffentlichungen werden von uns in der Rubrik Dagstuhl's Impact separat aufgelistet  und im Erdgeschoss der Bibliothek präsentiert.


Es besteht weiterhin die Möglichkeit, eine umfassende Kollektion begutachteter Arbeiten in der Reihe Dagstuhl Follow-Ups zu publizieren.