de

Welcome Gast

  • Login
Full load
SARD
Self-Management for Autonomic Relational Database Systems
Abstract
The featurism of commercial relational database management systems makes these systems applicable for a wide variety of applications. But their administration is a complex task and causes a high total cost of ownership for database systems. Autonomic (self-managing) database systems intend to reduce administration complexity and costs by automatically reacting to changing system loads and external events.
In the SARD research project we develop and evaluate techniques, which serve as a basis of a system-wide self-management logic for autonomic relational database systems:
  • Light-weight monitoring of the system state in order to detect the need for reconfiguration analysis
  • Modelling of DBMS components and behaviour as a basis for reconfiguration decisions
  • Definition and monitoring of business goals for autonomic databases
Participating members
  • Mitarbeiter
  • Studentische Mitarbeiter
    • Claas Gaidies
    • Okan Karakaya
    • Ali Haschemi
    • Sebastian Röhr
    • Sven Poggensee
    • Stephan Reinhardt
    • Andreas Kiolbassa
Publications of Project
2011
bibTexLogo
Marc Holze, Norbert Ritter
System Models for Goal-Driven Self-Management in Autonomic Databases
In: Data & Knowledge Engineering
2010
bibTexLogo
Marc Holze, Ali Haschimi, Norbert Ritter
Towards Workload-Aware Self-Management: Predicting Significant Shifts in the Workload
In: Workshop Proceedings of the 26th International Conference on Data Engineering
2009
bibTexLogo
Marc Holze, Claas Gaidies, Norbert Ritter
Consistent On-Line Classification of DBS Workload Events
In: Proceeedings of the 18th ACM Conference on Information and Knowledge Management
bibTexLogo
Marc Holze, Kathleen Krebs, Fabian Panse, Norbert Ritter
Data-Management-as-a-Service: Safe Ground or Terra Incognita?
In: Local Proceedings of the of the 13th East European Conference on Advances in Databases and Information Systems
2008
bibTexLogo
Marc Holze, Claas Gaidies, Norbert Ritter
Erkennung signifikanter Laständerungen für autonome Datenbanksysteme
In: Datenbank Spektrum
bibTexLogo
Marc Holze, Norbert Ritter
Autonomic Databases: Detection of Workload Shifts with n-Gram-Models
In: Proceedings of the 12 th East-European Conference on Advances in Databases and Information Systems
2007
bibTexLogo
Marc Holze, Norbert Ritter
Towards workload shift detection and prediction for autonomic databases
In: PIKM '07: Proceedings of the ACM first Ph.D. workshop in CIKM
Student theses within the project
2008
pdflogo
bibTexLogo
Diploma Thesis of Claas Gaidies
Klassifikation von Statements zur Erkennung von Workload-Shifts für Autonome Datenbanksysteme
Tutor: Marc Holze     Supervisors: Norbert Ritter, Axel Schmolitzky
pdflogo
bibTexLogo
Diploma Thesis of Björn Broll
Modeling, Prediction and Synthetic Generation of Database Workloads
Tutors: Marc Holze, Kathleen Krebs     Supervisors: Bernd Wolfinger, Norbert Ritter