Real time database - relative consistency

Real time database - Relative consistency, What is relative consistency in real time database, Explain relative consistency with an example

Relative consistency

A set of data items are relatively consistent when their values are updated within a certain time interval, called the relative validity interval (rvi).

A set of data items are relatively consistent if they are temporally correlated to each other. So, a derived data will be correct, only if the data items on which it is derived are relatively consistent with each other, or in other words, a set of data items would be called as relatively consistent, if they are contemporary data item.

Different data items that are used to derive new data must be temporally consistent with each other. This requires that a set of data items used to derive a new data item form a relative consistency set R.

Data item d is temporally consistent if and only if d is absolutely consistent and relatively consistent.

Example:

Assume that sensors are embedded in a car to monitor its fuel efficiency. They read the current fuel level and the distance travelled to derive the fuel efficiency. The car may go faster or slower. Depends on the speed there will be slight variation in the fuel consumption. And this variation may not go abnormal in any cases.

Let us assume that the current fuel level is 30 litres at time 100msec and the distance travelled is 10 kilo meters. Also assume that the avi is 20 msec. These real-time data can be specified as follows;

d_fuel (30 litres, 20 msec, 100 msec)

d_distance (10 kms, 20 msec, 110 msec)

We read these two values from two different sensors. So, there will be a difference or delay in data arrival at the controlling system. We have to accept the fact that both data are valid during this time delay. That acceptable time delay is called relative validity interval (rvi). Hence, if rvi is 15 msec for example, then the given data are relatively consistent. This time rvi is the difference between the time at which we read 30 litres and the time at which we read 10 kilo meters.

If the time difference between these two data is not more than the rvi value, then the data are consisdered to be relatively consistent.

 

***********************

Go back to Temporal Consistency page

Go to REAL TIME DATABASE Home page