As we discussed in the last post, there are number of ways to enforce access policies for any computing. Given the flexibility and dynamic nature of business, access policies should provide sufficient flexibility to meet business needs. To illustrate, let’s look at how an enforcement mechanism like network access control (NAC) can provide this kind of granularity. What you want is map out access models and design a set of policies to provide users with the right access at the right time from the right device.
Let’s focus on mobile devices, the poster children for any computing, and typically the hardest to secure. First we will define three general categories of mobile devices trying to connect to your network:
- Corporate devices: You have issued these devices to your employees and they are expected to get full access to pretty much whatever they need. You’ll want to verify both the user (strong authentication) and the device itself. It is also important to monitor what the device is doing to ensure authorized use after the pre-connect authentication.
- Personal devices: Sure, it’s easy to just implement a blanket policy of no personal devices. There are big companies doing that right now, regardless of user grumpiness over not being able to use their fancy new iPads at work. But if draconian isn’t an option in your shop, you could move authenticated, unauthorized devices onto a logical network configured only for outbound Internet access. Or provide access to non-critical resources such as employee wikis and the like but block access to corporate email servers, assuming you don’t want company email on these devices.
- Everything else: Lots of guests show up at your facilities and try to connect to your networks – both wired and wireless. If they successfully gain access via WPA2 or a physical port, they need to be bounced from the network. This represents the “access” part of network access control.
Depending on your pain threshold, there are many other device types and usage models that can be profiled to create specific enforcement policies. Granularity is only limited by your ability to map use cases and design access policies. Let’s not forget that you can also implement policies based on roles. For instance, your marketing group might have network access with iPads, since every good marketer needs one. But if engineers do not have a business justification for iPad use that group could be blocked. Policies aren’t defined merely by what (device) the user has, but also on who they are.
What about policies based on defenses implemented on the endpoint or mobile device – such as AV, full disk encryption, and remote wipe? Clearly you need to control those devices as well. Being able to restrict users without certain patches on their device is legitimate. Or you might want want to keep end users off of your protected network segment if they don’t have full disk encryption active, to avoid breach disclosure if they lose the device.
It’s not just about knowing what the device is, and who is using it, but also what’s on it. As you can see, this is problem includes at least 3 dimensions, which is why getting policies right is a prerequisite for controlling access. We’ll talk more about getting the policies right incrementally when we wrap up the series.
Which, once again, brings up our main point. Make sure you can enforce security policies that reflect your desired security posture given the context of your business processes. Don’t force your security policy to map to your enforcement mechanisms.