Going back many years, perimeter controls were traditionally adequate for protecting enterprise networks that held critical assets and data. The hypothesis was that if you had strong external perimeter controls, watching your ingress and egress should be adequate for protection. If you were a more sophisticated or larger entity, there would be additional fundamental separation points between portions of your environment. However these were still viewed and functioned as additional perimeter points, merely concentric circles of trust with the ability, more or less, to move freely about. In cases where threats occurred within your environment, you would hope to catch them as they crossed one of these rudimentary borders.

The Moment I Realized that Perimeters Aren’t Enough

This practice worked moderately well for a while. However, around fifteen years ago, security practitioners began to feel a nascent itch, a feeling that this was not enough. I personally remember working on a case, a hospital – attacked by a very early spear phishing attack that mimicked a help desk request for a password reset. Clicking on a URL in a very official looking email, staff were sent to a fakebut official looking website where these hospital professionals were prompted to reset their credentials – or so they thought. Instead, the attack began. This was before the days of the Darknet and we even caught the German hacker boasting about what he had done – sharing the phishing email and fake website on a hacker messaging board. I worked for a company that had a fantastic IPS solution and upon deploying it, we were able to quickly catch the individual’s exfils. At first, we seemed to be winning. We cut the attacker off from major portions of a botnet that resided on the cafeteria cash registers, most of the doctors machines and to my horror, even on the automated pharmacy fulfillment computers. Two weeks later, I received a call, the attacker was back,trying to get around the IPS device in new ways. While we were able to suppress the attack for the most part, I finally had to explain to the hospital IT staff that my IPS was merely at the entrances and exits of their network and that to really stop these attacks, we needed to look at all of the machines and applications that resided within their environment. We needed the ability to look at traffic before it made its way to and from the exits. This was to be the first of many realizations for me that the reliance on perimeter-based security was slowly and surely eroding.

In the years since, the concept of a perimeter has all but completely eroded. Of course, it took quite a while for the larger population to accept. This was helped along by the business and application interdependencies that bring vendors, contractors, distributors and applications through your enterprise as well as the emergence of cloud and cloud like provisioning utilized by Dev Ops. The concept of being able to have true perimeters as a main method of prevention is no longer tangible.

It was this reality that spurred the creation of Forrester’s Zero Trust model- almost a decade ago. The basic premise is that no person or device is automatically given access or trusted without verification. In theory, this is simple. In practice, however, especially in data centers that have become increasingly hybrid and complex, this can get complicated fast.

Visibility is Foundational for Zero Trust

A cornerstone of Zero Trust is to ‘assume access.’ This means that any enterprise should assume than an attacker has already breached the perimeter. This could be through stealing credentials, a phishing scam, basic hygiene issues like poor passwords, account control and patching regimen, an IoT or third-party device, a brute force attack, or literally limitless other new vectors that make up today’s dynamic data centers.

Protecting your digital crown jewels through this complex landscape is getting increasingly tough. From isolating sensitive data for compliance or customer security, to protecting the critical assets that your operation relies on to run smoothly, you need to be able to visualize, segment and enforce rules to create an air-tight path for communications through your ecosystem.

As John Kindervag, founder of Zero Trust once said, in removing “the Soft Chewy Center” and moving towards a Zero Trust environment, visibility is step one. Without having an accurate, real-time and historical map of your entire infrastructure, including on-premises and both public and private clouds, it’s impossible to be sure that you aren’t experiencing gaps or blind spots. As Forrester analyst Chase Cunningham mandates in the ZTX Ecosystem Strategic Plan, “Visibility is the key in defending any valuable asset. You can’t protect the invisible. The more visibility you have into your network across your business ecosystem, the better chance you have to quickly detect the tell-tale signs of a breach in progress and to stop it.”

What Should Enterprises Be Seeing to Enable a Zero Trust Model?

Visibility itself is a broad term. Here are some practical necessities that are the building blocks of Zero Trust, and that your map should include.

• Automated logging and monitoring: With an automated map of your whole infrastructure that updates without the need for manual support, your business has an always-accurate visualization of your data center. When something changes unexpectedly, this is immediately visible.
• Classification of critical assets and data: Your stakeholders need to be able to read what they can see. Labeling and classification are therefore an integral element of visibility. Flexible labeling and grouping of assets streamlines visibility, and later, policy creation.
• Relationships and dependencies: The best illustration of the relationships and dependencies of assets, applications and flows will give insight all the way down to process level.
• Context: This starts with historical data as well as real-time, so that enterprises can establish baselines to use for smart policy creation. Your context can be enhanced with orchestration metadata from the cloud or third-party APIs, imported automatically to give more understanding to what you’re visualizing.

Next Step… Segmentation!

Identifying all resources across all environments is just step one, but it’s an essential first step for a successful approach to establishing a Zero Trust model. Without visibility into users, their devices, workloads across all environments, applications, and data itself, moving onto segmentation is like grasping in the dark.

In contrast, with visibility at the start, it’s intuitive to sit down and identify your enterprise’s most critical assets, decide on your unique access permissions and grouping strategy for resources, and to make intelligent and dynamic modifications to policy at the speed of change.

Most people understand the reasoning and the reality behind a zero trust model. While historically, a network perimeter was considered sufficient to keep attacks at bay, today this is not the case. Zero trust security means that no one is trusted by default from inside or outside the network, and verification is required from everyone trying to gain access to resources on the network. This added layer of security has been shown to be much more useful and capable in preventing breaches.

But how organizations can move from a concept or idea into implementation? Using the same tools that are developed with 15-20 year old technologies is not adequate.

There is a growing demand for IT resources that can be accessed in a location-agnostic way, and cloud services are being used more widely than ever. These facts, on top of businesses embracing broader use of distributed application architectures, mean that both the traditional firewall and the Next Generation are no longer effective for risk reduction.
The other factor to consider is that new malware and attack vectors are being discovered every day, and businesses have no idea where the next threat might come from. It’s more important than ever to use micro-segmentation and micro-perimeters to limit the fallout of a cyber attack.

How does applying the best practices of zero trust combat these issues?

Simply put, implementing the zero trust model creates and enforces small segments of control around sensitive data and applications, increasing your data security overall. Businesses can use zero trust to monitor all network traffic for malicious activity or unauthorized access, limiting the risk of lateral movement through escalating user privileges and improving breach detection and incident response. As Forrester Research, who originally introduced the concept, explain, with zero trust, network policy can be managed from one central console through automation.

The Guardicore principles of zero trust

At Guardicore, we support IT teams in implementing zero trust with the support of our four high level principles. Together, they create an environment where you are best-placed to glean the benefits of zero trust.

• A least privilege access strategy: Access permissions are only assigned based on a well-defined need. ‘Never trust- always verify’. This doesn’t stop at users alone. We also include applications, and even the data itself, with continuous review of the need for access. Group permissions can help make this seamless, and then individual assets or elements can be removed from each group as necessary.
• Secure access to all resources: This is true no matter the location or its user. Our authentication level is the same both inside and outside of the local area network, for example services from the LAN will not be available via VPN.
• Access control at all levels: Both the network itself and each resource or application need multi-factor authentication.
• Audit everything: Rather than simply collecting data, we review all the logs that are manually collected, using automation to generate alerts where necessary. These bots perform multiple actions, such as our ‘nightwatch bot’ that generates phone calls to the right member of staff in the case of an emergency.

However, knowing these best principles and understanding the benefits behind zero trust is not the same as being able to implement securely and with the right amount of flexibility and control.

Many companies fall at the first hurdle, unsure how to gain full visibility of their ecosystem. Without this, it is impossible to define policy clearly, set up the correct alerts so that business can run as usual, or stay on top of costs. If your business does not have the right guidance or skill-sets, the zero trust model becomes a ‘nice to have’ in theory but not something that can be achieved in practice.

It all starts with the map

With a zero trust model that starts with deep visibility, you can automatically identify all resources across all environments, at both the application and network level. At this point, you can work out what you need to enforce, turning to technology once you know what you’re looking to build as a strategy for your business. Other solutions will start with their capabilities, using these to suggest enforcement, which is the opposite of what you need, and can leave gaps where you need policy the most.

It’s important to ensure that you have a method in place for classification so that stakeholders can understand what they are looking at on your map. We bring in data from third-party orchestration, using automation to create a highly accessible map that is simple to visualize across both technical and business teams. With a context-rich map, you can generate intelligence on malicious activity even at the application layer, and tightly enforce policy without worrying about the impact on business as usual.

With these best practices in mind, and a map as your foundation – your business can achieve the goals of zero trust, enforcing control around sensitive data and apps, finding malicious activity in network traffic, and centrally managing network policy with automation.