Covariance is a method of finding a relationship between two series.
Auto-covariance is used to find the variance of a series with itself, but the 2nd series start from the 2nd data point of 1st series (i. e. from 1-time lag).
It is used to find patterns in the series which might be hidden due to some noise or irregularity in the data.
Is covariance called irregular data?
We don't call covariance irregular data because it is not data, but is the result of the presence of relation in the data.