I read with horror TechCruch’s recent scoop that confidential business plans submitted to a prominent venture capital firm were accidentally published on the web. Ouch. It seems the database that the firm was using behind its online application was dumping the data into a human-readable backup file. The backup file lived in the same directory (folder) on their server as other web content, so it was accessible through a browser, and was indexed by Google.

Here are five tips for building any sort of web form or service that is going to handle confidential data:

1. Put your data on a different server (or servers) than your web files. The two don’t mix.
2. Lock down those database servers so only application code is allowed to connect to it, and the web servers are not.
3. Configure your backups to produce compressed, encrypted files, not human-readable files.
4. Move those backup files to tape archives. Did we mention not putting them on your web servers?
5. Turn off the Indexes option on your web server, which allows anyone (like Google’s automated crawlers) to see a list of all files in a directory on your server, even if they’re not linked to from a web page.
6. Bonus Tip: If you don’t know what you’re doing, go with someone who does — like, say, TrackVia!

• Previous Entry: Death of the Software Download
• Next Entry: Getting SaaS-y in a Downturn