Any merchant should protect their point-of-sale environment to the same level as the rest of their cardholder data environment. Maybe even a higher level since the POS's are more vulnerable to physical threats.
IBM Security Services PCI Blog
Andi Baritchi 27000216AA email@example.com Tags:  pci dss malware appliance anti-virus point-of-sale usb 708 Visits
If they see cardholder data - YES. Check out The TJX Effect and Keeping Point-of-Sale Equipment Secure for the rationale behind this.
Cute commercial from the PCI Security Standards Council explaining the virtues and requirements of the DSS...
Hat tip to Martin.
David Mundhenk 2700022KT6 firstname.lastname@example.org Tags:  pci nrf qsa compliance dss retail 1,839 Visits
It has been no secret that retail merchants and their industry representation have been extremely unhappy with having to comply with Payment Card Industry Data Security Standard (PCI DSS) requirements. Their industry representative presentations at the recent Congressional hearings on the effectiveness of PCI DSS showed that their spokespeople could barely contain their vitriol during testimony. In addition, the National Retail Federation has also recently published an ‘open letter’ to the PCI Security Standards Council regarding their collective discomfort over PCI requirements. A copy of this letter can be found at their own website via the following link:
This blog entry is not intended to make light of the concerns of the retail industry regarding PCI DSS, nor to dismiss their assertion that having to comply with the requirements isn’t pain or duty free. The challenges of complying with PCI DSS are measurable, and for some businesses difficult to address. This is especially true for small-to-medium sized businesses which may lack the expertise, bandwidth or resources to formulate an effective response to meet the requirements. The NRF collective attempts to make overall industry improvements for the better should be applauded, however, after reviewing the highlights of their recommendations it appears that they may have been drafted in haste.
What they have proposed is interesting and deserves some analysis and an objective response. The following is a representative sampling of NRF recommended improvements with respect to the current PCI compliance requirements, and also an interpretive viewpoint on the items noted:
National Retail Federation (NRF): Incorporate a formal review and comment phase on revisions to the PCI DSS by participating membership before they are issued.
Objective Response (OR): This is already taking place. The PCI SSC solicits feedback from certified QSAC’s and their respective QSA’s prior to finalizing version changes. In addition they sponsor “town hall” meetings where merchants and service providers who attend can ask questions, make recommendations, and vent their frustrations.
NRF: Ensure the amount of time from issuance of a revision to the
PCI DSS and the effective date is appropriate for all merchants, including
Level-1 merchants making enterprise-wide changes, based on the revisions that
are being implemented,
OR: This is purely subjective, no matter how much time is allotted there will be a contingent of merchants and others who feel like they need more time. The PCI SSC has to draw a line in the sand somewhere. In addition the PCI SSC has recently provided additional assistance for level 2, 3, and 4 merchants with their new, improved Self Assessment Questionnaires and their recommendations for a “prioritized” approach to achieving compliance.
NRF: Follow, and adopt, the ASC X9 announcement of its plan to develop a new standard to protect cardholder data that may include end to end data encryption.
OR: This is very interesting, and apparently their lone
recommendation toward relieving all the perceived pain and suffering associated
with having to comply with the PCI standard. They actually seriously propose
adopting yet another compliance standard? Especially one that is not even
mature yet? Yessir, and
They also propose adopting the supposed “holy grail” of cardholder data protection which is sometimes called “end-to-end” encryption. Sounds like a great idea in principle, but implementing end-to-end encryption will be a serious challenge for all parties currently involved in payment processing and settlement across the globe. It may actually be a viable prospect for the future and possibly successful over the long term. Initially, however, it will be incredibly expensive to deploy, administer, and maintain on a global scale. One also wonders what it will be like for everyone involved to re-initialize and then maintain all those crypto keys in a secure, PCI compliant fashion.
NRF: Utilize the concepts of key controls and controls rationalization to restructure the more than two hundred detailed requirements of the PCI DSS.
OR: This one too is fairly entertaining since the PCI DSS is already
constructed to accommodate what is known
SOX and SAS70 requirements permit unencumbered business operations to continue even if total compliance with their requirements has not been achieved. Also in some instances the auditor and the entity audited sit down prior to the exercise and “agree” upon the audit criteria to be measured against. "What? You don't want to rotate your encryption keys because you 'feel' like it is too much work? OK then! You don’t have to do it." This approach is not going work for PCI, which is one of the few required security frameworks that actually incorporates some level of enforcement for those who refuse to comply.
NRF: Require credit card companies and their banks to provide merchants with the option of keeping nothing more than the authorization code provided at the time of sale and a truncated receipt, rather than requiring merchants to store credit card information for dispute resolution, putting customers at unnecessary risk.
OR: Where PCI compliance is concerned this is a non-issue. Nothing within the PCI DSS nor the PCI PA-DSS requires the merchant to retain anything post authorization. Even section 3.1 & 3.2 of the PCI DSS clearly states the following:
3.1 Keep cardholder data storage to a minimum. Develop a
data retention and disposal policy. Limit storage amount and retention time to
that which is required for business, legal, and/or regulatory purposes,
3.2 Do not store sensitive authentication data after authorization (even if encrypted).
The PCI DSS doesn’t require having the merchants store any CHI post authorization. Some “issuer-processors” may actually be doing so on behalf of their customers (i.e. – banks, financial institutions, credit unions, etc.) who already own the data they are asking to be collected. In addition any sensitive authentication data being captured by them is usually collected during the pre-authorization process. Since the banking industry already "owns" this data, why shouldn't they be allowed to collect it? The additional audit compliance requirements imposed by the card brands on their issuer-processors accommodate a trusted relationship that allows for this to take place only under extremely special circumstances.
If a merchant does not want to collect CHI information for any reason what-so-ever post authorization they currently can choose not to do so. In reality it is usually the case that merchants have been collecting this type of information for years for their own business intelligence and data mining purposes. They are often unwilling to commit the resources required to re-tool their IT applications and infrastructure to support not storing CHI.
It has been reported in some Internet postings that the PCI DSS offers "minimum general recommendations for security" for systems storing or processing CHI. This is a myth. The PCI DSS is indeed a standard and a framework, not detailed process or procedural recommendations on how to secure PCI systems. This is somewhat akin to the OSI model defining the seven protocol and architecture layers for computing systems and networks, but not detailing “how” they should be configured down the system level. The corresponding PCI DSS Requirements and Security Assessment procedures, however, are exacting in spelling out how to measure PCI DSS compliance within a given environment.
The effects of changing or re-engineering an operational business enterprise to become PCI compliant are usually measurable and often significant. Sometimes dealing with any change to normal business operations especially in tough economic times is very uncomfortable, however, managing change and potential risk to business assets is a simply a major component of doing business safely and securely. Reducing or eliminating the emotional aspects of having to do so helps provide focus and reduces anxieties associated with doing so. It also helps to aid in complex decision making. The PCI DSS and related compliance processes are not perfect, but their adoption has resulted in a significant industry improvement over doing nothing, or adding additional security standards to the mix.
David Mundhenk 2700022KT6 email@example.com Tags:  secure dss compliance jfs pci delete 1,476 Visits
Meeting the Spirit of PCI Secure Delete
3.2 If sensitive authentication data is received and deleted, obtain and review the processes for deleting the data to verify that the data is unrecoverable.
The critical concept contained in this statement is that of rendering the sensitive data no longer needed “unrecoverable”. If sensitive data is stored within RAM then it is a pretty safe bet that once data has been deleted from RAM elements, then the intensive I/O dynamics of a system’s RAM offers its own built-in extensive data space ‘over-writing’ capabilities. In this sense the data is rendered “unrecoverable” very quickly, i.e. – no matter the time and resources available, it would be virtually impossible to reconstruct the original data structures. Even if one could, extracting and preserving real-time RAM data presents its own challenges and also should require highly privileged access to a given system. It is reasonable to conclude that the degree of difficulty in trying to extract bits and pieces of sensitive data from RAM given this scenario would render the data ‘virtually’ unrecoverable.
Securely deleting files from any file system including JFS requires using a ‘mil-spec’ data eradication utility. While doing so may not completely remove all sensitive data from low level data elements, it creates a scenario whereby a significant work effort is required try to recover meaningful data. It would require more so than if the data elements were simply deleted via file system utilities. Increasing the number of successive passes of this data eradication utility over the same data space would further reduce the probabilities that it could be successfully recovered.
So this yields the philosophical and yet highly practical question, “…how many passes from a ‘secure delete’ utility are enough to render the data virtually unrecoverable?” The answer to this question depends upon many factors, but obviously some reasonability should be applied here. The goal is to increase the difficulty and the amount of work required to recover any remaining data following the secure delete function, thereby rendering it an extremely time-consuming, expensive and highly dubious process. Ultimately the time and resources required to reconstruct such data would not be worth the prize of being able to successfully do so. In addition, limiting the types and quantities of sensitive data being stored and/or processed by the system also enhances this security proposition. This is because less sensitive stuff is ultimately being captured, preserved, and at some future time required to be ‘end-of-lifed’. Another consideration is to reduce the amount of journaling being done by the file system, and some JFS’ actually provide the capability to adjust journaling attributes. If possible, a combination of all of the above should be considered to optimize overall security and ensure compliance with PCI DSS requirements.
As most know there are relatively few absolutes in this world. Given this fact, Journaling File Systems present some interesting challenges with respect to PCI compliance. The ultimate goal is to provide a reasonable, cost effective effort to ensure that sensitive PCI information no longer needed be indeed rendered virtually ‘unrecoverable’. Now given the possibility of unlimited time and resources, it may be theoretically possible to recover data given the aforementioned scenario. The amount of time and effort required to do so, however, would cost far more than the data is worth. This yields an effective deterrent to those considering an attempt at such a data recovery. It would also require privileged access to the system, all of which should be detectable by the other required PCI controls for payment processing systems. As always the judgment on whether security controls are sufficient enough to validate PCI DSS compliance rests with the PCI Qualified Security Assessor.
Most working with PCI DSS and PA-DSS compliance are very familiar with the requirement to securely delete cardholder information and cryptographic materials once the data is no longer needed. Examples of such a requirement include:
3.2 If sensitive authentication data is received and deleted, obtain and review the processes for deleting the data to verify that the data is unrecoverable.
…and from the PA-DSS v1.2
1.1 If sensitive authentication data (see 1.1.1–1.1.4 below) is stored prior to authorization and then deleted, obtain and review methodology for deleting the data to determine that the data is unrecoverable.
This requirement is based upon the fact that simply deleting files from a Windows or Unix file systems does not completely purge the information from the hard drive or partition. During a delete operation, typically the metadata of the file is altered such as the file pointer, etc. has been altered, however, the actual data itself may still be resident in part or even whole. Currently, computer forensics professionals have various tools and techniques available to recover parts or all of the data contained with files that were simply deleted by the operating system. So the real question becomes how can sensitive data that is no longer needed be securely deleted, or in the very least, rendered such that is no longer ‘recoverable’ as described the above requirement from PCI DSS? To obtain some guidance regarding this issue we can review another section of the PCI PA-DSS v1.2:
1.1.4 Securely delete any magnetic stripe data, card validation values or codes, and PINs or PIN block data stored by previous versions of the payment application, in accordance with industry-accepted standards for secure deletion, as defined, for example by the list of approved products maintained by the National Security Agency, or by other State or National standards or regulations.
Now this sounds fairly straightforward; one simply obtains a commercial product or open-source software tool that has the capability to over-write consecutive binary data elements (usually all “1”’s or all “0”’s) over all the locations where the file contents were stored on the hard drive or partition. One directs the product or tool at the target files and enables a secure delete, “shred” or other function that consecutively over-writes the previously sensitive data numerous times. Problem solved, work is done, and case closed, right? Well, not necessarily.
While performing a PA-DSS assessment of the MainStreet Softworks payment processing application Monetra (which was ultimately found to be compliant by the way) IBM ISS PA-QSA’s were introduced to an interesting twist to the problem of secure file deletion. MainStreet Softworks CTO, and Senior Engineer Brad House revealed to the on-site QSA’s that it is extremely difficult if not even impossible to securely delete files from many modern file systems, especially what’s known as a Journaling File System (JFS). Brad has even written a technical paper addressing this issue in detail that even includes proof of concept code to prove this assertion.
What is a JFS? I am glad you asked. According to Wikipedia the definition of a Journaling File System is as follows:
“A journaling file system is a file system that logs changes to a journal (usually a circular log in a dedicated area) before committing them to the main file system. Such file systems are less likely to become corrupted in the event of power failure or system crash.”
Most, if not all Unix systems provide several options for file systems and many utilize Journaling. In addition most recent incarnations of the Windows Operating system also provide some JFS capability. As such, if systems that store, process, or transmit PCI cardholder information are built upon JFS systems, then attempting to securely delete sensitive authentication data and cardholder information becomes almost impossible. Even if one tries to implement some a tool or software technique to overwrite sensitive data within a JFS, there still is a probability that some, if not all of the data can be recovered using forensic tools and techniques.
So what can be done to be done to ensure the PCI compliance of credit card payment processing systems that are built on systems using JFS?
Next Post – Meeting the Spirit of PCI With Secure Delete
Andi Baritchi 27000216AA firstname.lastname@example.org Tags:  card compliance dss security credit iss cardholder pci 2 Comments 943 Visits
In today’s challenging economic landscape, clients in every vertical are looking harder than ever to lower costs and maximize ROI. It’s a delicate balancing act between tactical and strategic – one we often see left unbalanced.
In order to be successful, a security program must be:
All cost centers, including security and compliance, have been made prime targets for CFOs’ slash hunts. The most common mistake we’ve observed in such times is to shift security and compliance to a tactical-only paradigm. This shift may save costs in the short term, but chops ROI at the knees in the long run.
For example, building a security program for your organization wholly around PCI DSS compliance is a sexy, but fallacious idea. A security program built on top of the DSS is totally dependent on the Standard – you’ll have to re-think your security strategy every time the PCI Security Council issues an update (and/or clarification). Same goes for any other reg.
Before you ask us the question, I don’t have the budget to get an A in security – I just need a C in compliance, consider the consequences of building a security program that hyperfocuses on compliance.
To contrast, say you implement a sound security framework that, as one of its core dimensions, addresses cardholder data security and PCI DSS compliance. Other dimensions may address your company’s other regulatory, security, and privacy vectors – such as SOX404 compliance, web 2.0 threats, and so on. All the elements of this comprehensive security framework, if designed properly, will work together to keep your and your customers’ data secure. Compliance will just be a natural byproduct.
As a final thought, remember that any standard, whether it’s religious, regulatory, or simply an internal rule, can only be effective once all stakeholders understand and adopt it both in spirit and in function.