Network Segmentation: A Best Practice We Should All be Using

It would be nice to be able to say that we are winning the war; that network security efforts are slowly getting the better of the bad guys. But I cant do that. Despite all the money being thrown at security tools and hosted services, the cyber-thugs are improving their game at a faster rate than we are. The ten worst known cyber security breaches of this century have all taken place since 2008, and 2013 and 2014 are notorious for their information security incidents.

I think there are a multitude of reasons for this state of affairs to exist. One is confusion, indecisiveness and slow reaction times among regulatory bodies and standards providers. Another is the check the boxcompliance mentality that exists both in government agencies and in the private sector. A third is simply the insane rate of innovation in the information technology realm. There are many more. But despite the reasons, one thing is clear: we have to stop rigidly complying with baseline standards and move into the more flexible and effective world of best practices. And today the best practice I want to touch on is network segmentation.

In our business we see a lot of computer networks that are just flat. There is little or no network segmentation and anyone on the inside can pretty much see everything. I cant begin to tell you how easy this kind of setup makes it for us during penetration testing success is virtually assured! And its amazing how even just basic network segmentation can slow us down or stop us all together.

A good reason to start with network segmentation is that you can go at in easy stages. Maybe you can begin by segmenting off a separate development or test network. Those are pretty basic and can give your networking team some valuable experience for more difficult efforts to come. Then you can ensure that user spaceis separated from server space. Doing just that much can have an amazing effect – it really helps to thwart successful cyber-attacks.

As the team gains confidence in their abilities, they can move onto the next step: real enclaving of the network. This is anything but a trivial effort, and it requires detailed knowledge of the various functions of the different business departments and how information moves into and out of each one of them (a task made very much easier if the company has a good business continuity program and business impact analysis in place). But in the long run these efforts will be well worth the trouble. It is very difficult indeed to gain access to or exfiltrate information from a well enclaved network especially from the Internet.

This blog post by John Davis.

Touchdown Task for November- Network Segmentation Review

Whether it is budget preparation or annual project planning, the end of the year always leads us to think of the “big picture”. The touchdown task for this month is to review your network architecture maps and diagrams. First of all, make sure they are up-to-date. But secondly, look for indications that your network might be too flat. That is, do you have proper network segmentation between all of your information resources? Are your firewalls placed properly throughout your environment? 


A “flat” network architecture allows attackers who have gained a foothold on the internal (and sometimes even the external — you do have a layered DMZ, right?) network full visibility to internal systems and to move freely through workstation and server space. 


If you see some re-architecting that should be done, make note of it now. Depending on the complexity of the work, either schedule the re-architecture for a slow period at the end of this year or create a work plan for 2014. 

As always, thanks for reading and keep your eyes on the goal!

Three Examples of Thinking Differently About InfoSec

Today, I am putting my money where my mouth is. I have been talking about thinking differently about infosec as being a powerful tool in the future for several months now, but here are three concrete examples of how security folks need to think differently than they do today. (Note that some of you may have already begun to embrace these ideas – if so, awesome, you are ahead of the curve!)

#1 – Think like attackers AND defenders – We as infosec folks often get so caught up in our statements of ethics, credos and agreements about behavior that we get trapped inside them and become blind to the methods and ways of attackers. Many security folks I meet have taken such steps to distance themselves from attackers and they often show utter disdain for attackers, tools and techniques that they are essentially blind to the way attackers think. This is a dangerous paradox. If you don’t understand your opposition, you have no way of being effective in measuring your defensive capabilities. If you can’t think like an attacker, maneuver like an attacker and understand that they are not bound by the rules that you attempt to impose on them – then you will likely have little success in defending your organization against them. To better defend our assets, we have to be able and willing to understand our enemies. We have to have a realistic knowledge and capability to replicate, at the very least, their basic tools, techniques and attitudes. Otherwise, we are simply guessing at their next move. Essentially without insight and understanding, we are playing the “security lottery” in hopes of hitting the big defensive jackpot!

#2 – Deeper defenses are better defenses – We must extend defense in depth beyond an organizational approach to a data-centric approach. The closer to the data the controls are implemented, the more likely they are to be able to add security to the core critical data. (Of course, normal rationality applies here. The controls have to be rational, effective and properly implemented and managed – as always!) This is why security mechanisms like enclaving, data classification and eventually tagging are the future of enterprise security. If we start to think about our security postures, deployments and architectures with these ideas in mind today, we will be able to leverage them in their present state and eventually gain the maximum from them when they are fully ready for integration.

#3 – Think risk, not compliance – I am going to continue to talk about this, no matter how much heat I get from the “compliance guru set”. Striving for compliance with various regulations or standards is striving for the minimum. Guidance, regulations and law are meant to be the MINIMUM BASELINE for the work we need to do to separate liability from negligence.  Compliance is a milestone, not a goal. Effective understanding and management of risk is the goal. Don’t be deceived by the “compliance guru set’s” argument that meeting baselines if effective risk management. It is NOT. Regulatory compliance, ISO/PCI compliance pays little attention to and has little management for attacker techniques like vulnerability chaining, management/analysis of cascading failures or zero-day/black swan (Thanks, Alex!) evolutionary capabilities. This step requires upper management education and awareness as well, since those that control the budgets must come to see compliance as a mile marker and not the end of the race ribbon!

I hope this helps folks understand more about what I am saying when I assert than in 2008, we have to think differently if we want infosec to improve. Of course, thought has to precede action, but action is also required if we are going to change things. What is clear, from the problems of 2007 and further back, is that what we are doing now is NOT WORKING. It should be very clear to all infosec practitioners that we are losing the race between us at attackers!