How does True Image Verify an Image?

I was hoping one of the TI gurus here could enlighten me on how TI verifies an image... Does it actually do a "bit for bit" verification (if performed immediately after an image) or does it use some other method (such as CRC checks). Thanks for any responses!
howie

 

I'm not sure but think it counts the total files on both & see if the numbers match-not sure about that though.

 

Seekforever is the guru on this

As posted

"A validate is not a comparison of the archive contents with the source but rather a recalculation of the checksums and comparing their value with the original ones calculated and stored in the archive when it was created. TI writes 4000 checksums per gigabyte of data and every one must compare perfectly or else the archive is declared corrupt. This is a very rigourous test that demonstrates the archive can be properly read into RAM.
"

 

Thanks for that information, Tatou! I was pretty sure it wasn't "bit-for-bit", but I was not exactly sure how it was done... That pretty much explains it.
Thanks again!

 

Greetings:


howie123:

FYI a checksum or CRC match is the same as doing "bit-for-bit" but far more efficient. Checksums and CRC's are calculated at the byte or word level and the result still requires a "bit-for-bit" match in the data stream.

cheers

 

I believe he was using the term "bit-for-bit" to mean reading the archive and then comparing its contents with the original bytes of the source.

This would be possible (if implemented) using the TI rescue CD since it deals with a static source drive. It would be virtually impossible if making an archive while Windows is running because the source is changing all the time.

An advantage of comparing the archive contents with the source contents is the case where the data incorrectly arrives from the HD to the program for some reason. TI would make a checksum based on the incorrect data and from that time on, it would validate the archive correctly. The source drive comparison would likely catch such an error unless it was a repeatable error such as a data line stuck in one state. Fortunately, a repeatable error is likely to manifest itself in operations.

The checksum recalculation method also allows the integrity of the archive to be checked at any time in the future.