Segmentation Technologies / Zero Trust

This is an update to a post from 2001 which I’m revisiting in part because some things have changed, but also because (surprisingly) much hasn’t. 

I first came across the notion of doctrine vs. structure in this depiction about the relative positioning of tanks from a tweet I can’t now place. It has stuck with me for a while, not just because I’m interested in tanks, but rather because I really like this notion of thinking of doctrine (the intent of use, or overall philosophy of approach) as being something separate from structure (the thing you intend to use for that doctrinal purpose and how is it structured/built). I think we’d have a much better time in how we build security into our infrastructure or select the products we intend to use if we had more clarity of purpose in thinking about doctrine vs. structure - and to then look at how new adversarial approaches change the alignment. For example, how might the past few years of drone development and use in various theaters  reshape this chart.  

You can look across a whole set of spaces from technology to security and start to unpick the doctrine and structure, and position particular technologies on those axes. In doing so I find it forces you to think a bit harder about whether one set of features, technologies or products are being used in the right way. One example, below, is a chart for segmentation technologies (one element of, so called, zero trust architectures).

To be clear, I’m not saying this is complete or correct, it’s an illustrative example. But it is interesting to look at it this way and in particular look at what it doesn’t show. For example, I still can’t think of a good example of a doctrine purist and structure neutral technology in this context. Perhaps you can? Perhaps there shouldn’t be one? Or maybe there should be and this is in fact a new technology category. Maybe an example is the various technologies that bring VM and/or legacy infrastructure in as service mesh endpoints in a coarse-grained way. 

The other reason I like this approach is it does not mean that you only get to pick one technology to achieve the goal, in this case, segmentation in a zero trust context. Rather, it means you select a range of options to apply in the context of where that combination of technologies are meant to be used. Just like on a battlefield you get to pick a range of armored fighting vehicles to achieve a specific defensive or offensive outcome. But, unlike in armed conflict, we have more choice in our approach so we’re not confined to be “going into battle with the army we have”.

I can imagine developing this further to include an operational overlay with the hypothesis that if there is a fit between a use case and the technologies that balances the doctrine and structure for that use case then operational effort should be optimal.

Bottom line: thinking about doctrine vs. structure appears to be a useful mental model to validate a technology’s adequacy for a particular task. In short, to know whether we are jamming a square peg into a round hole.