Last updated at Fri, 29 Oct 2021 14:02:07 GMT
“Zero trust" is increasingly being heralded as the ultimate solution for organizational cyber safety and resilience — but what does it really mean, and how can you assess if it has a practical place in your organization's cybersecurity strategy for 2022?
In this post, we'll answer those questions by taking a look at what problems the concept of zero trust is trying to solve, what types of people, process, and technology are necessary for successful zero-trust implementations, and what mindset changes your organization many need to make to be fully ready for this new defender paradigm in the year to come.
What is zero trust?
At the core, the concept of zero trust is just what those two words suggest: every human, endpoint, mobile device, server, network component, network connection, application workload, business process, and flow of data is inherently untrusted. As such, they each must be authenticated and authorized continuously as each transaction is performed, and all actions must be auditable in real time and after the fact. Zero trust is a living system, with all access rules under continuous review and modification, and all allowed transactions under constant re-inspection.
What problems is zero trust trying to solve?
Zero trust aims to finally shatter the mythical concept of “castle and moat" (i.e., assuming individuals and components on the intranet are inherently safe) and fully realize the power of least privilege — the concept that individuals and components should only have the most minimal access necessary to perform a required action. We can see it better through the lens of a practical example, such as one of the most typical ransomware attack scenarios: an attacker gains initial access to a corporate network through simple VPN credentials.
In most current implementations, a VPN has one interface that sits on the internet and one that sits on the intranet. Unfortunately, most VPNs are still accessed via simple credentials. Once authenticated, an attacker impersonating a user represented by those credentials has general network access. They're free to replay the credentials (or attempt to use various tools to obtain other credentials or tokens) on any other connected system until they gain access to one where they can elevate privileges and begin exfiltrating data and corrupting the integrity of filesystems and databases.
In a zero-trust environment, the user identified by a set of credentials would also need a second authentication factor. The entire authentication attempt would be risk-assessed in real time to see if the individual's connection is, say, in an allowed geofence and that the access time is within the usual operating mode of that person (and that the individual does not already have an established session).
Even if an attacker managed to obtain multi-factor codes — for example, SMS 2-factor authentication (2FA) has weaknesses but may be the only 2FA an organization can afford to implement — they may achieve a successful connection but would not have general access to all intranet systems and services. In fact, the VPN connection would only grant them access to a defined set of applications or services. If the attacker makes any attempt to try a network scan or perform other noisy network actions, monitoring systems would be alerted, and that individual and connection would be quarantined for investigation.
Each transaction has a defined set of authentication, authorization, and behavior auditing rules that continually let the overarching zero-trust system ensure the safety of the interactions.
What do you need to move to zero trust?
While this section could fill an entire book, we'll work under the assumption that you are just beginning your zero-trust journey. To make this initial move, you'll need to pick at least one business process or service access scenario to move to this new model.
Every component and individual that is responsible for enabling that business process or service must be identified and the architecture fully documented. At this point in the process, you may find that you need to reimagine the architecture to ensure you have the necessary control and audit points in place. You'll then need authentication, authorization, auditing, risk-assessing, and enforcement solutions to support the access decisions at each connection in the process or service. Finally, you'll need staffing to support creation and maintenance of the rules that are enforced, along with traditional patching, mitigation, and configuration management enforcement activities.
Then, lather, rinse, and repeat for all other processes and services. In other words, you need quite a bit.
However, you should not — and, in reality, cannot — move every business process and service to zero trust all at once. Once you've assessed that initial service, begin the groundwork of acquiring the necessary tools and hiring the necessary staff to ensure a successful outcome. Then, transition that initial service over to zero trust when funding and time are on your side, and leave it in place for a while as you evaluate what it takes to maintain safety and resilience. Adjust your tooling and staffing plans accordingly, and get to work on the remaining processes or services.
Thankfully, you may have many of these components and personnel in place within existing security and compliance solutions and processes, and you can finally employ more of your existing investments' capabilities than the 5 to 15% that most organizations generally utilize.
Adopting the zero-trust mindset
One of the biggest mindset challenges to overcome when introducing zero trust into your organization is the fear that the constraints the model imposes will reduce productivity and hamper creativity. These fears can be overcome with the right framing of zero trust.
Start by performing a scenario-based risk assessment of a given business process. Do this with the business process owner(s) or stakeholder(s), and ensure you enumerate what actions threat actors could take at each transaction point in the process, ideally with some measurement to the costs due to loss of safety and resilience.
Then, show how each threat is reduced or eliminated with a zero-trust implementation of the same business process, and note how new processes — developed with a zero-trust mindset at the start — will have reduced implementation costs, be far more safe and resilient, and be much easier to enhance over time as they will have been established on a solid foundation.
Zero trust is not some sticker on some point solution's brochure. It is a fundamental change to how your organization approaches access, authentication, authorization, auditing, and continuous monitoring. You won't adopt zero trust overnight, but you can begin that journey today, knowing that you're on the path to helping your organization protect itself from tomorrow's threats, as well as today's.