Software Quality Insights:

PCI compliance

Feb 5 2009   2:20PM GMT

Does PCI compliance make your data secure? Nope.



Posted by: Jack Danahy
PCI compliance, software security

Another week, another cascade of information pouring unintentionally out of another unwitting company — this time it is Heartland Payment Systems.

As a result, Heartland customers will get letters letting them know that they should watch out for unexpected transactions; hundreds of man hours are going to be spent understanding the circumstances of the breach; and already the inquisitors of information security are pounding keyboards mercilessly, pillorying the Heartland team for this most recent episode.

Heartland is reported to have been PCI-compliant. That’s the interesting nugget in this story and it’s one we have seen before. PCI compliance didn’t save Heartland from losing so much data that the company could more aptly be called ”Heartland Pay-out Systems” for the next several months, as it pays out clean-up, fines, and costs.

Assuming Heartland did what was asked by PCI auditors, and that it provided sufficient access for the assessor to do their job: if I were Heartland I would be really not happy. If there was some fundamental communication breakdown, I would be similarly displeased.

I mean, really, what is the purpose of issuing one of the most proscriptive standards on security if measurement of compliance with it is so meaningless? My information about this data breach comes from public sources and not from Heartland or their PCI assessors, but here is what I can glean:

1. Heartland was certified as PCI-compliant. (See the list of PCI-compliant service providers. Heartland is on Page 12.)

2. Part of PCI compliance includes section 3.4 which says that the credit card data must be encrypted anytime it is stored. Now, while some will argue that there should be a new requirement calling for encryption of internal networks (which does not exist in PCI currently), any entry-level programmer knows that as soon as I read my input for the credit card number, I am storing it, even if only in memory. (See the complete PCI DSS v1.2 standard.)

3. Malicious code — likely a sniffer – grabbed private unencrypted data (they called it “recorded data” which sounds a lot like like storage to me) off of the wire as the data was being sent for processing. (See here for one of many articles about the breach.)

So, Heartland wasn’t compliant, period, according to my read; but note that I am not a certified reviewer.

Is this their fault or their assessor’s? I have no idea, but I do believe that it is the responsibility of the assessor to have done this simple analysis and identified the weakness.

If Heartland was complicit in this, then I have no defense for them. If, however, in a business climate charged with specialization and outsourced expertise, they relied on their provider to help them validate that they had done enough, then the responsibility and the spotlight should be turned at least equally on the process and the people who gave them the useless rubber stamp.

Does being PCI-compliant make you secure? Nope. Not the same thing at all, but that isn’t the issue here.

Does thinking you are compliant when you are not even close create some substantial risk?

You bet.