QRecall 4 times faster than filesystem copy

FYI, just a little note to let you know some real world performance results.

I use QRecall to backup my Mac, including a 25GB data file for a VMWare Fusion Virtual Machine. Today, I selected the VMWare data file in QRecall and asked it to Recapture just that file. It had been about a month since I last backed it up, and the operation took 6.5 minutes.

Then I decided to copy the VMWare data file from my Mac to the same external drive that QRecall uses. I wondered if the operation would be faster or slower (I was guessing that it would be faster to do the file copy than to use QRecall because I suspected that most of QRecall's performance comes when incrementally backing up lots of small files, and I thought it might not do so well on a single large file).

I was surprised to find that the file copy operation took 25 minutes! I was pleased that even when backing up a huge binary data file, QRecall's ability to identify the bytes that have changed and only copy them makes a huge performance impact.

Nice work!

Michael Trafton wrote:Nice work!

You're welcome.

I was surprised to find that the file copy operation took 25 minutes! I was pleased that even when backing up a huge binary data file, QRecall's ability to identify the bytes that have changed and only copy them makes a huge performance impact.

The biggest performance advantage, when capturing large amounts of duplicate data, is that the read performance for most drives is significantly faster than the write performance. When capturing duplicate data, QRecall spends most of the time reading from the archive. This is (ironically) faster than just blindly writing all new data.

Thanks for the performance report!