Zero-trust is an increasingly popular cybersecurity model. Even the National Security Agency encourages the use of a zero-trust architecture, largely because of its data-centric approach to protecting critical assets across the network. Yet, no matter how good it sounds, it isn’t a perfect solution, as the NSA also points out.
“Systems that are designed using zero-trust principles should be better positioned to address existing threats, but transitioning to such a system requires careful planning to avoid weakening the security posture along the way.”
That said, despite the many clear advantages of using zero-trust, how trustworthy are zero-trust architectures?
The NIST Tenets of Zero-Trust
There is an NIST framework of zero-trust architecture that includes the following:
• Data and computing services are considered resources.
• All communication should be secured, no matter its network location.
• Access to resources should be on a per-session basis.
• Access to resources can be determined through identity and/or behavior attributes.
• Monitoring security posture of resources.
• Enforce authentication and authorization.
• Use data collection to continue to improve security posture.
The framework is just that—a framework to begin building zero-trust architecture.
“These tenets are useful guideposts, but simply not enough to fully guide zero-trust architecture and subsequent implementation,” said Andrew Smallwood, chief technologist with Booz Allen Hamilton, speaking at the (ISC)2 Security Congress.
How Zero-Trust is Untrustworthy
Despite the increasing popularity and increasing adoption of zero-trust architecture, Smallwood warned there are pitfalls that could result in devastating cybersecurity incidents.
“Zero-trust is not a silver bullet. You can’t just set it up and forget about it,” Smallwood said. It requires vigilance and monitoring, and it also opens up your network and data to other security issues that must be addressed, which include:
• The assumption of full visibility of assets and control of network actions. Knowing your assets has to come before zero-trust is deployed, not after, or you will miss out on valuable protections.
• Dealing with legacy systems. Almost every organization has some legacy architecture in their security stack. Legacy systems hinder zero-trust because they don’t have the capability to support modern segmentation or identity and access management (IAM). Replacement of legacy systems is complicated; not only is it expensive, but in many cases, these systems perform their functions better than newer technologies, especially within government agencies.
• Updating across technologies is budget- and time-intensive. Zero-trust requires reclassifying authorization and database updates. The security architecture also needs to mesh smoothly with any digital transformation initiatives within the organization.
• It might be the wrong security architecture for OT systems.
• Zero-trust may not protect against shadow IT and BYOD. There are too many workarounds users have discovered that will allow them to bypass zero-trust architectures.
Downgrading Cybersecurity Posture
If these issues aren’t considered while transitioning to a zero-trust architecture, it could end up hurting your overall cybersecurity posture.
One problem that Smallwood has seen is a self-inflicted DDoS attack, where the organization making the transition to zero-trust unintentionally limits access and controls.
“By not properly establishing them, the organization could limit access to resources,” Smallwood said.
Also, the use of shadow IT and the workarounds used to access it creates a larger attack surface and then hampers the organization from fully using the full complement of its security resources.
There are two solutions available to resolve the problems within zero-trust, according to Smallwood—governance solutions and technology solutions, as well as hybrid solutions combining the two. Governance solutions involve long-term leadership engagement to ensure compliance, working groups to oversee implementation and to address the different stakeholder needs.
Technology solutions will allow organizations to take action, such as automatically denying access unless the user is approved as authorized and using tech offerings like app-aware firewalls to make sure segmentation is done correctly.
Hybrid solutions offer a holistic view of the zero-trust architecture and strategy. Zero-trust architecture can be extremely effective when the NIST framework is used as a foundation and the architecture is built with potential security problems in mind. If not, it can actually weaken an organization’s security posture.