Measuring and Managing Information Risk: A FAIR Approach

In a world where seemingly everything is oversold, this is the rare exception that is undersold. The title succinctly states, without drama, the authors’ broad ambit. They over-deliver. The book is nothing less than a manifesto for quantitative management of information security risk. Consider how radical it is to promise a truly quantitative approach to cyber risk management in a world dominated by numerous qualitative “frameworks,” red-yellow-green heat maps, thousand-item one-size-fits-all questionnaires, subjective and qualitative scales of likelihood and impact, and fake math like “red times green equals yellow”. And then consider how transformational it is to deliver on the promise. Other reviewers have nicely discussed the book’s coverage of the FAIR taxonomy. Suffice it to say that MMIR is your best friend in understanding the Open Group FAIR standards. Freund and Jones bring a potentially dry subject alive with many “Talking About Risk” sidebars that tell of their experience with FAIR methods in practice. These war stories make the content accessible and relevant. I especially appreciate the authors’ informal style that is conversational without being verbose and humorous without being patronizing or cute. What the war stories leave out chapter 8 fills in with numerous example analyses. A worked example is better than a thousand war stories. If giving a thorough rationale for and introduction to FAIR were all that MMIR did, it would be worth its weight in gold. But wait! There’s more! It’s the “managing” part, chapters 11-14, that constitutes another breakthrough beyond FAIR. There Freund and Jones begin laying out (one senses it is a work in progress) a risk management ontology, built on the FAIR risk measurement ontology. In rethinking the classification of controls in the context of threat event frequency, vulnerability, and loss mitigation, they provide ways to assess and – yikes! – quantify the potential value of control improvements, in isolation or in combination. This gives the CISO the beginning of a way to manage the control environment, not just the threats. But controls not consistently adhered to are both false comfort and all too common. Therefore F&J suggest that variance in the application of controls is perhaps the single most important set of infosec management metrics. As the old saw goes, if you cannot measure it you cannot manage it, and if you do not know how well your controls are operating on a continuing basis, then what confidence can you have in the millions of dollars invested in technology and staff? Which brings us to metrics. It is perhaps not surprising that a methodology based on quantitative analysis lends itself to meaningful metrics. F&J offer many concrete suggestions far superior to the grab-bag of metrics found in vendor dashboards (measure what’s cheap and looks cool) and other books. These are real metrics that the CISO can use to … manage risk. And managing risk is really why we do all this stuff. Making good decisions on both operational and strategic levels requires good data derived from reliable instruments and methods. It is in managing risk that MMIR is truly seminal and profound. If they do another edition Freund and Jones should consider adding a subtitle, “The CISO’s Bible,” because CISOs will find themselves coming back to it time and again. Or maybe that is the next book.

Measuring and Managing Information Risk: A FAIR approach is not like “traditional” information risk texts. Freund and Jones have taken the method far beyond what we are used to. They have the audacity to question the all too familiar “likelihood” and “impact scores” as well as their product, the “risk value” and its canvas, the "risk matrix". The premier contribution they give is the amount of thought that went into their method and ridding it from illogical aspects. This has made the method somewhat contrived in my opinion but you can’t escape the fact that it fits together. Also, the authors put a big emphasis on using a careful and consistent terminology, something that is truly bothersome in information security risk today. The FAIR method for analyzing risk does have rigor and may prove to give reliable results. Question is: will anyone but a select few bother to use it? Regardless, the book is a tremendous introduction to serious information security risk analysis, you will scoff at risk values and matrices after reading it.

This is a great way for a information security professional to turn away from the "soft" methods of heat maps and ordinal scales. Information security needs real probabilistic methods to solve real risk assessment problems. Jack Jones and Jack Freund have found a great way to introduce these concepts by building a practical probabilistic method that cybersecurity experts can use - and already have used. I highly recommend it as required reading for information security professionals at all levels. Douglas W. Hubbard Co-author of How to Measure Anything in Cybersecurity Risk

I'm rather familiar with FAIR, and its revision by the OpenGroup: https://www2.opengroup.org/ogsys/jsp/publications/PublicationDetails.jsp?publicationid=12239, https://www2.opengroup.org/ogsys/catalog/C13K https://www2.opengroup.org/ogsys/catalog/C13G https://www2.opengroup.org/ogsys/jsp/publications/PublicationDetails.jsp?publicationid=12158 This book does not duplicate the existing literature on FAIR, but goes into the specific details of how FAIR is used and the algorithms involved in the specific steps. Unfortunately, there is no attempt to explain the construction of Monte Carlo simulations, recommending the CXOWare solution (expensive) or OpenPERT https://code.google.com/p/openpert/ (free plugin for MS Excel). FAIR relies heavily on Monte Carlo simulation. The volume spends time to teach the differences between frequency and probability analysis, and the traps of both. It teaches the ontology that the OpenGroup has also published, as well as terminology specific to FAIR. It shows how to defectively measure It discusses how to calibrate measurements, how to deal with the limitations of probabilistic models and how to handle issues of accuracy versus precision. It also provides an excellent guide to interpreting the results, and the common mistakes analysts make. The chapter on controls is well thought out as it shows how to breakdown the overlap between prevention/detection/response in a control set, as well as how to understand the impact of the fact that all controls are vulnerable to some degree, and therefor only effective at a percentage. Unfortunately, there is no good data that experts can use to calibrate these values, as none of the various breach reports expose the failed control set. Please note that this gap is not a failure of the book, and I'm just raising my favorite complaint. The chapter on metrics is somewhat obvious. Quantitative risk analysis produces measurements and those measurements can be compared with goals to allow for metrics. They make the correct (in my estimation) to focus on impact here. If you want to learn how to leverage a very serious quantitative analysis tool, this book is well worth the purchase. The book itself is produced by a print on demand service, and has some font issues, where the font is often rather small and hard to read for my old eyes. Paper quality is high, so the book is rather heavy for its thin binding. Other than this, if you are considering a quantitative method for performing risk analysis, I can highly recommend this volume. A somewhat critical note on FAIR as a methodology The use of Monte Carlo simulation for risk analysis is well documented as a successful approach, but relies upon the problematic PERT distribution. PERT has not been shown to be mathematically valid, and has arbitrary input shapes. (Ferson & Shoemaker). PERT, however, has the advantage of allowing the mathematical capturing of that calibration through an adjustment of the variables. Another mechanism to capture the calibration of measurement mathematically is a p-box. It would be interesting to try to build a monte carlo simulation built upon a p-box instead of PERT. Unlike PERT, p-boxes are mathematically valid and allows you to marry intervals with probability, distinguish between variability and incertitude and like PERT allows you to work with unknown input distributions.

good book

Superb book. I bought it after seeing Jack speak in Sydney. His talk immediately "clicked" in my mind, connecting a lot of dots, that in retrospect were obvious.

Great way to break down InfoSec Risks into tangible impacts and provide credible articulations of risk. A much needed move away from pure prescriptive best practice controls (implicit risk management) to focused controls to assets that matter most. (Explicit risk management).

Reading the book is a start, but it is not really helpful. Yes, certain information are supplemented to the free material that can be found on the internet. Here are my main points, why the book is just partially useful: 1. Ok, it is a good point that it is not a book supposed to teach Monte-Carlo-simulations. I get this part. There are good study materials that do the job. But, and this is the big but, there is no explanation or any kind of useful information on how the variables interact from a mathematical viewpoint. Knowing this would allow to program it yourself, for example with python or r. Of course, understanding the math behind Monte-Carlo-simulations and how they work, it is possible to come with a solution (your solution). But that is my point, you must came up with the (your) solution, it is no shown or descripted in the book. Therefore you cannot be sure whether you “solution” on how you have interpreted the FAIR model is somewhat right or not. 2. The book is just to “hook” you up. In order to go further you need the FAIR training. Many information bits are missing in the book, preventing you to actually using it. I know that because of my “aha” moments during “a product” presentation by a certain company, offering there FAIR “Tool”. During the presentation, certain aspects on how to use FAIR become clearer. Reading the book is not helpful to actually use FAIR in a useful manner. In order to do that (well guess what, yes that’s right!) you need the offered training, provided (guess again) by the same company, that is behind FAIR (yes a company, check out who is behind the training and FAIR, too!). The book is therefore only partially useful, more to hook you up, even if you do understand Monte-Carlo-simulations, so do not expect too much from the book.

The. FAIR methodology is an important part of the toolkit for Risk Management practitioners. This book does a superb job of introducing FAIR as a quantitative assessment methodology. Among other things, it deals with the practical reality of doing risk assessments in a large organization and communicating results to executives.

A great book! I bought this after hearing Jack Jones speak at a conference in October, and it has changed the way I look at risk, for our own company and those we help