Self-managing database systems

1
Self-managing database systems

• Erik Gast
• Stijn de Gouw
• Johannes Kruisselbrink

2
Self-managing database systems

• Minimize user-interference with the underlying
  system of the databases.
• Properties of a self-managing database
• Transparent underlying system
• Optimize data-structures and internal structures
  with respect to the usage of the database
• Detect and repair missing data, erroneous data
  and anomalies

3
Why is this so hard?

• It difficult to keep the underlying system in an
  optimal state while the data and the usage of the
  system are changing
• Optimizing operations with respect to usage
  requires accurate predictions about the future
  usage of the database system
• Semantic knowledge is required to detect and fix
  errors and anomalies

4
Why is this useful

• With a self-managing system, people don't need to
  bother about the underlying system anymore and
  can solely focus on the data
• Maintaining a database becomes easier
• Reduced user interaction with the database
  management reduces the costs
• Better performance of database operations
• etc.

5
Is this testable?

• Yes, because
• Performance of the system can be measured.
• User-involved maintenance work can be measured.
• But testing the self-managing database with
  different kind of data is hard.

6
Incremental

• Plenty of intermediate milestones
• Self-management can be divided into a lot of
  smaller research areas that all contribute to the
  degree of self-management.
• Each of these sub-parts of self-management
  already exist, but could be improved.
• i.e. better error detection 4, more reliable
  error-recovery 4, easier database maintenance
  1, 2, 3, better operation optimization 1
  w.r.t. the usage of the database system.

7
References

• 1 Oracle Database 10g, The Self-Managing
  Database, an Oracle White Paper, Nov. 2003
• 2 T. Nakajima, H. Aizu, M. Kobayashi, K.
  Shimamoto "Environment Server A System Support
  for Adaptive Distributed Applications", In
  Proceedings of the International Conference on
  World Wide Computing and its Applications'98,
  1998
• 3 E Gjørven, F Eliassen, K Lund, VSW Eide, R
  Staehli, International Workshop on Self-Managed
  Networks, Systems , 2006 heim.ifi.uio.no,
  Self-Adaptive Systems A Middleware Managed
  Approach
• 4 Hurson, A. R., Jin, B., and Pakzad, S. H.
  1991. Neural network-based decision making for
  large incomplete databases. In Proceedings on
  Parallel Architectures and Languages Europe
  Volume I Parallel Architectures and Algorithms
  Volume I Parallel Architectures and Algorithms
  (Eindhoven, The Netherlands). E. H. Aarts, J. van
  Leeuwen, and M. Rem, Eds. Springer-Verlag New
  York, New York, NY, 321-337