Error rate of Data Transfers?

[Don J]

I notice that when transferring very large amounts of data the copy display
starts by bouncing around by a factor of about three in the amount of
remaining time, with most of the time about 75% spent on the lower values.
This leaves about 25% of the time at the larger values. I've always assumed
that it is monitoring the error rate and that fluctutations in error rate is
what it is displaying. I assume that what it is designed to do when error
rate is high is to repeatedly retry the transfer. If correct my
observations implies that there is a considerable amount of cut and try.
The error rate on the cut and try must be significant.

Am I right? What is the logic ot the data transfers? I am particularly
interested in the value of the residual error rate. Can anybody help me?

Don J

--------------------------------------------------

 

[Pegasus \(MVP\)]

I notice that when transferring very large amounts of data the copy display
starts by bouncing around by a factor of about three in the amount of
remaining time, with most of the time about 75% spent on the lower values.
This leaves about 25% of the time at the larger values. I've always assumed
that it is monitoring the error rate and that fluctutations in error rate is
what it is displaying. I assume that what it is designed to do when error
rate is high is to repeatedly retry the transfer. If correct my
observations implies that there is a considerable amount of cut and try.
The error rate on the cut and try must be significant.

Am I right? What is the logic ot the data transfers? I am particularly
interested in the value of the residual error rate. Can anybody help me?

Don J

Your assumption is most likely incorrect. The display jumps about
because the estimate is based on file numbers rather than file sizes.
When the copy process hits a large file, the estimate jumps up. When
it deals with a batch of small files, the estimate drops because the
copy process appears to progress very rapidly.