eraser vs file shredder

silver163

New Member
I noticed a difference between eraser and file shredder in the way work. I used a same file size and type (encrypted container generated by truecrypt) but eraser was a lot quicker than file shredder so i am worried that if eraser is going quicker it may not be as secure? both programs use the 35 pass algorithm.
 
How much faster are we talking about here?
 
How much data are you erasing? If File Shredder takes 1 minute, a one minute difference makes Eraser infinitely faster; if File Shredder takes 1 hour, then I'm inclined to say that Eraser's speed difference can be attributed solely to experimental error. Perhaps a percentage and absolute amount of time would be needed to have an accurate picture here.

Are you using a multiple-pass erasure method?
 
Joel said:
How much data are you erasing? If File Shredder takes 1 minute, a one minute difference makes Eraser infinitely faster; if File Shredder takes 1 hour, then I'm inclined to say that Eraser's speed difference can be attributed solely to experimental error. Perhaps a percentage and absolute amount of time would be needed to have an accurate picture here.

Are you using a multiple-pass erasure method?
I generated two containers, exactly 500 MB each, using same algorithm, same password and used 35 pass algorithm to erase both files. The difference between them was a minute. You can try it out too, it just requires Truecrypt and Eraser and File Shredder.
 
I think that seems fine, mathematically at least.

I wouldn't know how to convince you that Eraser is secure, short of letting you read the source code? The passes are as Gutmann described in his paper. The only passes which deviate from the official specification are those which require verification. It would not be practical to write one pass then verify that the pass was properly written, if the pass was pseudorandom data.
 
Joel said:
I think that seems fine, mathematically at least.

I wouldn't know how to convince you that Eraser is secure, short of letting you read the source code? The passes are as Gutmann described in his paper. The only passes which deviate from the official specification are those which require verification. It would not be practical to write one pass then verify that the pass was properly written, if the pass was pseudorandom data.
I guess I just really want to know why the difference though if the same file and same algorithm?
 
I'm curious to know the difference too, but I can't think of anything at this point in time.
 
Joel said:
I'm curious to know the difference too, but I can't think of anything at this point in time.
I think it be better if Eraser showed which pass it is currently working when erasing a file.

Also one additional thing I noticed, when a file is for some reason somehow in use (in my instance it was an .iso file) Eraser doesn't inform you that is in use, it simply crashes.
 
When the task is running, double click on the task, Eraser reports which pass is being executed.

The crash when a file is locked has been fixed in 6.0.9 (this affects Vista SP2, I'm not sure if SP1 or RTM is affected. Windows 7 is not affected.)
 
Joel said:
When the task is running, double click on the task, Eraser reports which pass is being executed.

The crash when a file is locked has been fixed in 6.0.9 (this affects Vista SP2, I'm not sure if SP1 or RTM is affected. Windows 7 is not affected.)
This occurred with 6.1.0.2284 on Windows 7 Pro 64 bit
 
6.1.0.2284 is an older build than 6.0.9, when the nightly build server is back up the new builds should have the fix.
 
Joel said:
6.1.0.2284 is an older build than 6.0.9, when the nightly build server is back up the new builds should have the fix.
really? thought that 6.1 is newer than 6.0 but ok I will switch to 6.0.9
 
Well, 6.1 will be newer than 6.0 when it is released. Both are in parallel development; in the mean time, to know which has the latest fixes, see the last (large) number. The larger number is the newer build.

The first two numbers tell you which branch it is. 6.1 is the development branch where new features and breaking changes are planned. 6.0 is the stable where only bug fixes are applied.
 
Back
Top