| Home | Highlights | Intro | Objective | System | More |

| General Issues And Design Goals | Basic System Components |

General Issues

Distributed Intelligence

It is common knowledge that there is a lot of number-crunching involved in the evaluation of seismic data. How could a primitive low power controller in the RED System´s front-end perform such a demanding task?

First, the controller is not as primitive as one may think. This device can be operated economically in a low-power mode during simple data acquisition, but it can be sped up to a clock rate of 8 MHz (equivalent to 2 MOPS) on demand—whenever more computing power is necessary.

Even more important: the RED System´s performance strongly depends on distributed processing. In a real earthquake, there are at least hundreds, maybe thousands, of stations involved in the evaluation of individual raw data sets. This means that all these stations are busy at the same time with the evaluation of different seismic data sets—all of which belong to the same event. A massive parallel computing architecture!

The mathematics are simple: the total computing power of the RED System during a major earthquake reaches an incredible performance figure: several giga-operations per second. And there is no mass-storage involved in the front-ends, so the operands can be fetched within a fraction of a microsecond.

Such massive parallel computing outperforms even most of today´s workstation computers. Due to the high degree of redundancy in the stations, the parallel architecture has a unique robustness. The RED Systems Network comprises distributed intelligence at its best!

And each of the front-end units has sufficient battery backup capacity to continue operation for several days!