Design Methodology for Allocating Data in a Distributed Database.
Abstract
The objectives of this research was to find a method to solve the problem of deciding the best means for allocating data in a distributed data base. A decision function which considered system response times of update and query messages, storage cost of data items, and execution cost for allocating data was proposed. Additionally, the decision function allowed weighting of decision criteria to tailor the selection to specific network conditions. Five data allocation methods were evaluated for a given test network. The allocation methods were automated to facilitate future research efforts. The test network was implemented as a simulation model consisting of six nodes and eight data items. The simulation model was both verified and validated. Statistical analysis was performed on selected outputs of the model. Normalized data from the simulation model, normalized output of the automated allocation methods and scenarios of weighted decision criteria for certain network conditions were evaluated by the decision function. The utility of the decision function was demonstrated through comparison of the results of each scenario. The impact of the research results were discussed and areas of future research were presented. Keywords: Tactical data system, Communications network distributed databases, Simulation, Information systems, Command and control, Theses.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1987
- Accession Number
- ADA188826
Entities
People
- Edward T. Poore Jr
Organizations
- Air Force Institute of Technology