StorefrontBacktalk

PCI's Not-So-Open Global Forum

Written by Stephen Ames
July 22nd, 2013
Stephen Ames, CISA, CISSP, is the Director of Information Security at Shift4, a payment vendor.

PCI's Global Forum is an open forum in name only, at least as long as it continues to force changes on members that they are not permitted to even know about until someone who has been briefed chooses to tell them. What makes me say that? Let me tell you a story about how PCI really works.

Just wrapped up onsite PA-DSS validations with my PA-QSA this month and a question came up about PA-DSS Requirement 4.2.7, which aligns with DSS Requirement 10.2, which is all about user access. Just so we're all on the same page, you need to know that none of my company's PA-DSS applications have a user database. Hence, all of PA-DSS Requirements 3 and 4 are not in scope for us. That's the way it has been since Visa's PABP days and beginning with Version 1 of the PA-DSS.

However, my PA-QSA stated that PA-DSS Requirement 4.2.7 is now always in scope, regardless of whether or not there is a user database within the application. He went on to say that Reports on Validation (ROVs) used to be accepted by the PCI SSC with Requirement 4.2.7 marked "N/A" in the absence of any user database, but the SSC apparently recently "reinterpreted" Requirement 4.2.7 and sent guidance out to the assessor community in some newsletter. More on that later.

He gave me a few options to satisfy his checkbox, which actually correlates with PCI DSS 11.5:

  • Provide detailed instructions in the PA-DSS Implementation Guide on how to perform object-level auditing with all supported operating systems on all program executables and static files associated with the application.
  • Bolt on a COTS or internally developed hashing utility and automatically run periodic CRCs.

    Both of these options would cause application vendors (my employer included) to take on more liability. I've searched the PA-DSS for a security requirement that aligns with PCI DSS 11.5 — File Integrity Monitoring — and there is none. I'm certain that most application vendors would not take responsibility for file integrity monitoring at merchant sites. And I'm unable to understand why the SSC is forcing that upon application vendors, when they don't even have that requirement written into the PA-DSS.

    I searched the PCI FAQ database and found no reference to a reinterpretation of PA-DSS Requirement 4.2.7 requiring vendors to take responsibility for file integrity monitoring of their PA-DSS applications running in merchant environments. Once again, PA-DSS Requirement 4.2.7 aligns with DSS Requirement 10.2 and user access, not DSS Requirement 11.5.

    So, back to that newsletter that I keep hearing about every time I have a dispute with a QSA or PA-QSA. My question is always, "What newsletter?" and the response is always, "The SSC sends out compliance guidance to the assessor community."

    Really? Are we going to have this argument again, PCI Council?Really? Are we going to have this argument again, PCI Council? I thought we resolved this back in Orlando in 2010 when I took the open mic on the subject at the PCI Community Meeting. At that time it was about "additional DSS reporting requirements" that I knew nothing about until my QSA was on site explaining to me why he was asking for a lot more detail for his executive summary and checkboxes. At that meeting, the Council admitted they needed to do a better job of keeping the entire community informed of important changes to the Data Security Standards and avoid suppressing important updates from participating organizations. In fact, the very next day, the additional reporting requirements were posted in the Documents Library. But it now appears the PCI SSC has fallen back into its old ways of keeping participating organizations in the dark.

    I just can't figure out why I have to learn about security standards changes or so-called "new guidance" from my QSA or PA-QSA and not via official correspondence from the PCI SSC to all participating organizations. It makes no sense to tell only the assessor community. Why is there a secret society within a self-proclaimed, open global forum?

    I reviewed the lifecycle policy for PCI DSS and PA-DSS, and there is a provision for "minor changes or errata" the Council may publish at any time. Unfortunately, the Council publishes these changes in a newsletter that goes out only to the assessor community. Is this where new standards, revisions, and errata — that apply globally — should be published? I think not.

    In its own words, the PCI SSC defines itself as "an open global forum." Open global forum? What part of this practice is open or global? What the Council needs is a global newsletter that goes out to the entire PCI community so that we can all be on the same page.

    Now it's been a few years since I last took a political science course and studied the Rule of Law, but I do recall a specific requirement that all laws be publicized before they are enforced. And that publication of the law has to be broad enough so that any interested party can find it. Well, here I am. I am an interested party, and I want to know more about this change to Requirement 4.2.7, but I can't find any written information. Until I do see officially published documentation, I will consider the PCI Council's guidance to the assessor community regarding PA-DSS Requirement 4.2.7 to be an illegal change to an existing, in-force security requirement.

    You need to fix this, PCI SSC. So please come down from your ivory tower and follow your own written policies. I'm tired of the PCI SSC using the QSAs to bully participating organizations. From personal experience, I know that many QSAs and PA-QSAs are frustrated that they constantly have to explain the Council's position in the absence of any concrete, black-and-white security standard. That is just wrong.

    I look forward to PCI's response to and remedy of this situation. I welcome comments, questions, and commiseration. And to those assessors drinking the PCI Kool-Aid, please don't clutter the Internet with rebuttals or holier-than-thou rhetoric. There is no PCI DSS 11.5 alignment within the PCI Payment Application Data Security Standard, and until there is, you have no leg to stand on. Period.