The diverse platforms of edge computing pose a challenge when attempting to consolidate them into a single security stack. Consequently, networks become vulnerable to unforeseen endpoint attacks. However, edge and IoT platform providers have recently shifted away from the "trust but verify" philosophy and have started incorporating technology that treats every endpoint and identity as a new security perimeter.
The truth is that most of today's edge and IoT platforms were not designed with sufficient security measures to withstand endpoint attacks. Companies struggle to integrate these platforms into a unified security stack because legacy edge and IoT platforms rely on server and operating system security. The lack of enforcement of least-privileged access through interdomain trust relationships leaves a wide range of endpoints vulnerable to intrusion and breach attempts.
To prevent devastating breaches, companies need to secure edge computing and IoT platforms across their entire stack. This includes hardware, operating systems, application platforms, data, and network security. Companies must explore how zero-trust can address the challenge of securing complete technology stacks for edge computing and IoT networks.
Zero-trust is described in the National Institute of Standards and Technology Special Publication (NIST SP) 800-207 as a new approach for trusting IT resources.
![Zero-trust data points](/static/media/9198938f-8c47-5a0e-82d9-6db6a62cd3f7/DAM-b09bb69a-e771-4560-8b30-a28ab48636be/out/2=GUID-23D778D3-FE3C-473E-A1C2-A6AFBF8C2A30=1=en-us=Low.png)
Zero-trust is an integrated, end-to-end security strategy based on three core principles:
- Never trust, always verify—Always authenticate and authorize based on all available data points, including user identity, location, device, data sources, service, or workload. Continuous verification means there are no trusted zones, devices, or users. Instead, zero-trust treats everyone and everything as a potential threat.
- Assume breach—By assuming your defenses have already been infiltrated, you can take a stronger security posture against potential threats, minimizing the impact if a breach does occur. Limit the "blast radius"-the extent and reach of potential damage incurred by a breach-by segmenting access and reducing your attack surface, verifying end-to-end encryption, and monitoring your network in real time.
- Apply least-privileged access—Zero-trust follows the principle of least privilege (PoLP), which is the practice of limiting access rights for any entity and only permitting the minimum privileges necessary to perform its function. In other words, PoLP prevents users, accounts, computing processes, and so on, from having unnecessarily broad access across the network. Such access leaves your network vulnerable and creates a higher attack surface in case of a breach.
John Kindervag developed the original zero-trust model in 2010, and as a principal analyst at Forrester Research, he realized that traditional access models operated on the outdated assumption that organizations should trust everything within their networks. The thinking was that perimeter-based security (in other words, firewalls) would be enough to validate user access and secure the network entirely. But as more workers started remotely accessing systems through all types of devices and all kinds of connections, this trust structure proved insufficient to effectively manage a distributed workforce. Kindervag recognized this vulnerability and developed zero trust in response.
A few major tenets:
- Network segmentation—Traditional networks exposed direct access to all data assets, servers, and applications. The zero-trust model segments various subsets of these resources and removes the ability for users to directly access them without first going through a tightly controlled gateway. This is sometimes referred to as “network isolation.” Micro segmentation takes this concept further by isolating workloads from one another so that administrators can monitor and control the flow of information between different servers and applications rather than just between client and server.
- Access control—Regardless of whether users are physically located in an office or working remotely, they should only be able to access the information and resources that are appropriate for their respective roles. Each segment of the network should authenticate and validate authorization to ensure that traffic is being sent from authenticated and authorized users regardless of the location or source of the request.
- Visibility—Network boundaries should inspect and log all traffic, and admins should regularly monitor logs to ensure that users are only attempting to access systems that they are permitted to access. Administrators often use cloud access security broker software to monitor traffic between users and cloud applications and warn when they see suspicious behavior.