Going Out With A Bang

2010 ended with a slew of database breaches being reported. Mozilla's database for addons.mozilla.org was hacked, Ohio State University leaked 760,000 student and faculty social security numbers, St. Louis University reported a breach, and Silverpop leaked most of its clients' customer email addresses.

But in the greater scheme of things, these breaches will have the same financial impact as CitySights NY's leaking 110,000 credit card numbers.

CitySights NY had 110,000 credit card numbers, along with both personal information (name, address, email) and CVV2 data -- sometimes called the security code. Storage of the CCVV2 is expressly forbidden under these circumstances by the Payment Card Industry Data Security Standard. It stands to recon that if you have 110k credit card numbers, you might have heard of PCI-DSS, or have been flagged by a PCI assessor. Somehow CitySights NY got around this restriction.

Each one of these cases involves a database breach, caused by anything from SQL injection attack to simple user errors that expose data to anyone who happens to stumble upon it. It's easy to become desensitized reading about breaches given the frequency has barely slowed in the past five years. Most companies storing credit card numbers have received the memo that they need to take care of their data, but there are a lot that have lax security and are still storing portions of the mag stripe data that they should not. I find that universities remain defiant about data security, and most institutions I speak with have a cultural issues that run counter to data security. An open environment that promotes the free exchange of ideas is at odds with the need to lock down personal information. It's not that they can't -- it's just counter to their philosophy, and I believe why we will continue to see student data compromised for the foreseeable future.

Whatever the cause, we have technologies and processes to help mitigate breaches and make is far tougher for attackers to compromise a database. What's not always obvious is that a single account compromise or SQL injection attack should not automatically expose every data record, or all types of data. It does not need to be catastrophic and total failure.

Simple adjustments to permissions, segregation of admin roles, or patching more frequently don't cost money -- rather it's a change to the operations processes. What's more is these steps account for some of the biggest breaches. Labeling and masking technologies, commonly available from the database vendor at a small cost, can prevent leakage through mistakes and some forms of injection attacks. Transparent database encryption is, once again, inexpensive and effective with minimal disruption to operations. Finally, choosing to _not_ store some types of information is free and incredibly effective.

You don't need to do a lot to raise the bar, but you actually have to _do_ something. Outside of Mozilla, it does not look like any of these institutions had security in mind whatsoever.

Adrian Lane is an analyst/CTO with Securosis LLC, an independent security consulting practice. Special to Dark Reading.