Jacob Nikom
Track: Data Warehouse and Data Marts
Date: Tuesday, April 19
Time: 4:50pm - 5:35pm
Location: Ballroom G
TrackBack
The Reagan Test Site Operations Coordination Center (ROCC) manages the data
flow from multiple sensors, processes the acquired data, and provides operator
displays to track and predict the path of space objects.
Recognizing whether two sensors see the same or two different objects is a major challenge. The decision-making process at ROCC uses various data-mining techniques based on analysis and comparison of past and current measurements. Efficient implementation requires that ROCC have a data warehouse architecture integrated with powerful data analysis tools.
Another challenge is effectively processing data streams that come from multiple sensors. Features that are not typical of the traditional database management system, such as almost real-time response, high input data rates, specific structure of incoming data, etc., validate the need for a sensor
database.
In addition to these often contradictory requirements, the data collection process is very intensive. It involves many well-trained operators and costly high-precision instrumentation. Lost data due to database failure is inexcusable. Reliability is paramount in the selection of a storage system, narrowing the choice to a relational database such as MySQL. In addition, the design must take every precaution to prevent the database to became a single point of failure for the whole system.
This talk discusses the type and efficiency of relational database architecture that will satisfy the preceding requirements.
Download presentation file