Posts tagged threats
When it comes to PCI-DSS, it is easy to get confused about whether or not it’s working. And part of the reason why is that it has never been very clear what problem the PCI-DSS is attempting to solve. Is it trying to prevent fraud, or ensure a dependable minimum level of security in the payment system? My answer so far is neither.
Fraud has always been a problem of payment systems. Cards, like cash, can be counterfeited – and as technology to make counterfeiting more difficult advances, so too does the technology with which anti-counterfeiting methods can be defeated. In card payments, liability for fraudulent transactions is defined within their operating rules (for example the Visa Operating Regulations), and tends to be determined on a transaction-by-transaction basis. To prevent fraud the card issuer that authorizes the transactions needs as much information as possible to detect off behavior, and the merchant that accepts the transaction needs to take some basic precautions at the point-of-sale (in the U.S., at the simplest level this is swipe the card and check the signature). Liability for fraud in the face-to-face environment (when the merchant follows the correct operating procedures) usually rests with the Issuing bank. Liability for fraud in the Card Not Present (CNP) world often rests with the merchant, because the merchant *can’t* follow the existing procedures — no signature.
The point here is that liability is determined on a case-by-case basis, applying the operating rules to the details of each transaction, as evidenced by the data that has gone back and forth between the merchant and cardholder, and then from the merchant through their acquirer/processor to the issuing bank and back again. Transactional liability is both well-defined and, given the scenario, relatively easy to assign.
However when payments started going online, something interesting happened. It became fairly obvious that the information needed to process a payment online (16-digit PAN, expiration date, address information of the cardholder) was also (obviously) being transmitted online and (not so obviously, but, in the early 2000’s kind of terrifying) being stored online. This opened up the possibility that an entity could get popped and lose not just a week’s worth of transactions at one point-of-sale — but hundreds, thousands, *millions* of cards in a single swoop, and fraudsters could use those cards downstream.
Let’s review that scenario: an online retailer gets popped and then those cards get used…at OTHER online retailers. Or face-to-face retailers that allow key-entered (manually typed in) transactions. Maybe across many retailers. And across many Issuers. Not a “local” merchant, like a gas station, where it would be fairly obvious to connect the fraud cases that follow. How long would it take to detect it? How many downstream participants in the system, following the operating procedures as designed, would have to deal with the negative aftershocks coming from that one compromise event? And since the party that “should” be accountable is not actually part of the manifesting fraud transactions, how can liability be shifted?
Historical note (fraud prevention infrastructure): In the early 2000’s the discussion was mostly focused on the CNP environment because a) it was new, b) it was the wild west, and c) a number of high profile web companies got hax0red. Payment/fraud geeks can think of this as the era of: Address Verification System (AVS) was *pretty* well established at this point, SET was finally acknowledged as totally DOA, the drumbeat for 3D Secure was on but *nobody* was adopting yet (it was before the big push in the EU)…and early days for chip. Also: still on regular DES in the PIN infrastructure.
Historical note (fraud prevention & security strategy): One may also remember the era in this way: Issuing banks owned authorization strategy (meaning, they were making the approval decisions on transactions) and very few merchants had made investments in fraud screening. All of the banks were getting a little spooked that databases full of juicy card details were sitting outside the payment system and that so many of them were accessible from the internet. Merchants of yore never needed to store card details — just receipts.
Back to the scenario: one of those wild west outposts full of juicy card data gets popped, downstream participants (Issuers, merchants, cardholders) feel the pain, the compromised party may or may not be known. The network operating guidelines’ transactional rules don’t adequately assign liability back to the accountable party: what are the banks going to do? Well, there are pretty much two options: adjust the system to assign liability back to an accountable party OR go outside the system to demand restitution. The former is difficult and the latter takes issues of the payment system outside standard channels which for several reasons is not ideal for the payment systems themselves, who have elaborate systems set up for arbitration and compliance to address issues between participants.
It is out of this alchemy that the card network data protection programs were born. They are liability plays all the way, and I give them the benefit of the doubt that they were meant to be incentives to encourage merchants to “do the right thing” and secure payment card data. MasterCard and Visa developed slightly different programs. My paraphrase is: MasterCard’s SDP essentially said to merchants — we trust you to secure the data, but if you get hacked we are going to levy fines to high heaven. And Visa’s CISP/AIS programs essentially said — we want to see some proof in advance — so get audited by a Qualified Security Assessor (QSA) who will attest you’re cool, and then if something happens but you were compliant, we’ll work with you.
Both approaches are sticks, neither is a carrot. The subsequent merger and morph into the PCI-DSS sort of also merged the compliance program approach. Merchants must both get attestations of compliance AND if breached there are programs for providing remuneration to downstream system participants affected. (BTW: PCI has more than one standard in it’s purview, though most people when referring to PCI mean the Data Security Standard…poor PIN/PED requirements, always subordinated to DSS…)
The PCI-DSS requirements themselves were developed as a set of reasonable industry best practices (I can only speak w/authority on intentions behind the Visa program progenitors but let’s just go with it). At their best, they are meant to provide guidance to merchants who otherwise would have no idea how to protect cardholder data. The criticisms of the DSS itself are wide-ranging but I personally find the DSS simply basic, not bad, and just a narrow view (focused on payment card data and systems) and yet still very general (security is contextual and needs to be tailored). Really the DSS should have been left as guidance, but unfortunately there’s this business of being assessed and a whole industry that has grown-up around QSA-dom. I find the process of getting assessed to be much more objectionable than the requirements themselves: “compensating controls” could be a book in and of itself, but QSA’s are auditors first and foremost (ROC on), and generally asked to interpret or design strategic security as a secondary concern, if at all.
Now, while in the early 2000’s the focus of all this angst was on CNP merchants, since then the scope of the PCI-DSS problem has embiggened. The DSS quickly expanded to include non-CNP, payment processors, and even the banks themselves. So a program of best practices designed to secure payment card data *outside* of the payments infrastructure got some retrofitting to also cover payment card data ostensibly *inside* the payments infrastructure. You can’t see me right now, but I’m raising an eyebrow, because the payment infrastructure is so interdependent, and has so many legacy components – to step-level up the security of payments infrastructure is impossible to manage without some serious planning, a boatload of direct economic impacts, and hell of a lot of specificity. *Upgrade to triple-DES I’m looking at you*. All I’m saying is, if you’ve got payments infrastructure requirements – don’t bring a knife to a gun fight.
Speaking of knife fights, let’s chat about fraud. Yes, fraud may go up after a major compromise. If counterfeiters flood the market with bad cards, it may take a while for the issuer fraud screening to kick in. Card re-issuance is expensive and so some issuers take a risk and leave potentially compromised cards open and then miss some fraud transactions. However, if major compromises go down, will the fraud rate also go down? That is less clear. Motherlode-sized compromises are a recent phenomenon, and while the fraud rates have ticked-up in the past two years, from 2003-2010 they were hovering near historical lows. Note: fraud losses in total dollars continue to climb, but the global fraud rates (i.e. the portions of total payment volume that ends up as fraud) are relatively steady over the past decade, less than 6 basis points (percents of a percent). (If you’re a U.S. CNP merchant you’re laughing out loud that people would be in a huff over 6 bp). The Nilson Report is a good place to get some data. I also like the CyberSource report but I’m writing this all in notepad.
My point here is that major compromises (that have been getting a LOT of press and attention) are only ONE method the fraud economy uses to operate. All of other methods like skimming, social engineering, insider threats, and plain old theft still exist. And so with all of this, how is fraud being kept down below 10bp? That’s a multiple choice answer: some markets have opted for prevention strategies (Europe loves their chip & PIN and 3D Secure is working there), others have opted for more advanced detection strategies (in the U.S., both issuers and merchants have adopted more advanced fraud screening technology). There are a lot of influences, but it’s pretty clear that most entities that get hit with transactional fraud losses are a) not waiting around for a panacea, and b) not depending on upstream security to reduce their exposure to fraud. (Fraud counterpoint: if you’ve got fraud prevention requirements, don’t bother with a gun in a knife fight.)
Thus if we are to ask ourselves what the PCI-DSS program (requirements plus compliance program) is set up to solve, the answer is something along the lines of “to provide a benchmark of *NOT negligent*” for individual system participants. And that might actually be an okay scope, as long as everyone’s clear what problem is being solved and that it is in the industry/community’s best interest to solve it. However, to solve problems like “fraud prevention” or “payments infrastructure security”, stronger — or at least more direct — medicine (and economic incentives) will be required.
This particular post was inspired in part by this Business Week Online article. As an industry, if we are looking to make improvements to infrastructure security or fraud management, we need to be asking the right questions. And as we seek to improve defenses and system strategy in general, it’s useful to clarify the different problem spaces of fraud & security, if only to confirm the variety of solution sets (technology, process, economics, compliance) we have to work with.
Risk modeling, while it sounds specific, is actually super-contextual. I think my own perspective on the topic (the different types of modeling, what they are good for) was best summed-up in a paper/presentation combo I worked on with Alex Hutton for Black Hat & SOURCE Barcelona in 2010. Probably video from Barcelona is the best reference if you want to look that up (yes, lazy blogger is lazy), but let me summarize by the (from my perspective) three general purposes of risk models:
- Design: Aligned most with system theory. The models try to summarize the inputs (threats, vulns, motives, protections) and the outputs (generally, loss and in some cases “gains”) of a system, based on some understanding of mechanisms in the system that will allow or impede inputs as catalysts/diffusers of outputs. Generally I would lump attack tree modeling and threat modeling into this family, just a different perspective on a system as a network architecture or design of a protocol, software, or network stack. Outside of risk/security, a general “business model” is equivalent, which attempts to clarify the scope, size, cost,and expected performance of the project.
- Management: Aligned most with the security/risk metrics movement, and (to some extent) aligned with “GRC”-type work, management-focused risk models are set-up to measure and estimate performance, i.e. to answer a questions about “how well are controls mitigating risk” or “to how much risk are we exposed”. One could think of the output of the design phase being a view as to what types of outcomes to expect, and then the management phase will provide a view on what outcomes are actually being generated by a system/organization. Outside of risk/security, a good example of a management model is the adoption of annual/quarterly/ongoing quality goals, and regular review of performance against targets.
- Operations: Operational models are a different beast. And my favorite. Operational models aren’t trying to describe a system, they are embedded into the system, they influence the activities taking place in the system, often in real-time. I suppose any set of heuristics could be included in this definition, including ACL’s. I prefer to focus on models that take multiple variables into consideration – not necessarily complex variables – and generate scores or vectors of scores. Why? Because generally the quality of decision (model fit, accuracy, performance, cost/benefit trade-off) will be more optimized, i.e. better. Outside of risk/security, a good example is dynamic traffic routing used in intelligent transport systems.
“Framework” is another term that I’ve heard used in a number of different ways but it seems to really be an explanation of a selected approach to modeling, and then some bits on process – how models and processes will be applied in an ongoing approach to administer the system. Even Wikipedia shies away from an over-arching definition, the closest we get is “conceptual framework“: described as an outline of possible courses of action or to present a preferred approach to an idea or thought. They suggest we also look at the definition for scaffolding: “a structure used as a guide to build something” – (yes, thank you, I want us to start discussing risk scaffolding when we review architecture, pls)
A little blog post.
So, it’s been about two years since I added anything to this blog. I’ve been busy!! The awesome folks at SOURCE gave me a speaking slot at SOURCE Boston 2010 and that kicked-off a series of talks on methods consumer-facing companies/websites take to protect customers from online threats. And then later in 2010 was able to participate in some discussions on different types of threat modeling and situations in which modeling techniques can be useful.
In 2011 I wanted to talk about some more concrete topics, and so spent some time researching how threats/impacts can be better measured. This is an area I’d like to spend more time researching, because there’s still a gap between what we can do with the the high-frequency/lower-impact events (which seem to be easier to instrument, measure, and predict) and the lower-frequency/high-impact events (which are very difficult to instrument measure, or predict). –> I think the key is that high-impact events usually represent a series or cascade of smaller failures, but there’s more research into change management and economics to be done.
Later in 2011 I switched over to describing how analytics can be used to study and automate security event detection. I hope in the process I didn’t blind anyone with data science. (haha…where’s that cowbell?) So here’s what I did: (more…)