Modern enterprise applications are increasingly built using containerized microservices orchestrated by Kubernetes. This architecture delivers significant advantages including scalability, resilience, portability, and faster software delivery, but it also introduces a substantially larger attack surface. Security must extend across multiple layers, including application code, container images, Kubernetes configurations, cluster services, and the underlying infrastructure. A single misconfiguration or compromised component can expose an entire application environment.
Traditional Kubernetes security relies on defense-in-depth mechanisms such as Role-Based Access Control (RBAC), network policies, admission controllers, image scanning, runtime monitoring, and service meshes. While these controls are essential, they ultimately assume that the underlying host operating system, Kubernetes node, and infrastructure can be trusted. If these foundational layers are compromised, sensitive workloads and application secrets may still be exposed.
By leveraging hardware with Confidential Computing capabilities such as AMD SEV-SNP and Intel TDX, cloud-native applications can finally enforce hardware protection directly to containerized workloads. This new approach ensures that application memory remains encrypted and isolated throughout execution, protecting sensitive workloads even from privileged infrastructure administrators.
This enables organizations to securely deploy modern applications across public cloud, private cloud, hybrid, edge, and sovereign cloud environments while significantly reducing infrastructure trust assumptions.
Key Security Benefits
Confidential Computing for containerised applications provide several capabilities that strengthen Kubernetes security beyond traditional controls:
- Protection against privileged infrastructure access, preventing cloud providers or system administrators from inspecting application memory.
- Isolation from host-level compromises, limiting the impact of attacks targeting the host operating system, hypervisor, or Kubernetes node.
- Protection in multi-tenant environments, preventing neighbouring workloads from accessing sensitive application memory.
Hardware-backed remote attestation, ensuring that applications only execute after verifying the integrity of both the infrastructure and the software stack. - Protection of data in use, extending encryption beyond storage and network communications to the processing phase itself.
Together, these capabilities establish a zero-trust runtime where trust is based on cryptographic verification rather than assumptions about the underlying infrastructure.
From confidential containers to trusted Kubernetes deployments
While Confidential Computing capability provides hardware-enforced protection, organizations also need an operational mechanism to ensure that sensitive workloads are deployed only on verified infrastructure.
This is where CanaryBit Surveyor extends the Confidential Computing model.
Surveyor acts as a confidential pod provisioner for Kubernetes environments, ensuring that containers and pods are scheduled only after the underlying infrastructure has been cryptographically verified by CanaryBit Inspector, an extended remote attestation service.
Instead of assuming that a Kubernetes cluster node is trustworthy, Surveyor validates the integrity of the execution environment before sensitive workloads are executed. This transforms Confidential Computing from a hardware capability into an enforceable runtime security policy.
Surveyor supports two deployment models, allowing organizations to balance security, isolation, and performance according to their workload requirements:
Kata mode for maximum isolation (Recommended)
In kata-mode, every container or pod is launched inside a lightweight confidential virtual machine using Kata Containers.
Before the workload starts, CanaryBit Inspector performs remote attestation to verify the integrity and characteristics of the underlying infrastructure.
This approach provides:
- Hardware-enforced isolation for every container.
- Strong protection against container escape and host compromise.
- Independent security boundaries for each microservice.
- Cryptographic proof that workloads execute in a trusted environment.
Kata-mode is particularly suited for regulated industries, multi-tenant platforms, confidential AI, and applications processing highly sensitive data.
Node-mode for wider-availability and higher performance
For workloads where performance and density are priorities, Surveyor can launch containers directly on confidential nodes.
In this mode, the node itself is remotely attested by CanaryBit Inspector before workloads are deployed, ensuring that applications only execute on trusted infrastructure.
This model provides:
- Protection of data in use through Confidential Computing.
- Hardware-backed verification of the Kubernetes node.
- Lower virtualization overhead and improved performance.
While containers share the same confidential node and therefore do not benefit from the additional isolation provided by Kata Containers, this approach is well suited for trusted enterprise clusters running on standard Kubernetes deployments.
Bringing zero-trust to Cloud-Native applications
As organizations modernize applications into microservices and deploy AI workloads on Kubernetes platforms, securing only the infrastructure is no longer sufficient. Trust must extend to every layer of the technology stack and every deployment decision.
By combining:
- Confidential Computing for hardware-protected execution,
- Remote Attestation through CanaryBit Inspector for continuous verification,
- Surveyor for trusted workload orchestration and policy enforcement,
CanaryBit delivers a suite of products for Kubernetes confidential deployments.
Organizations can finally both deploy sensitive workloads securely and verify digital infrastructure before sensitive data, encryption keys, or AI models are ever exposed.
This enables enterprises to confidently deploy sensitive applications and mission-critical microservices across cloud, hybrid, edge, and sovereign environments. Furthermore, it helps reduce infrastructure risks and support the security objectives of regulations such as GDPR, the EU AI Act, DORA, and NIS2.
In a zero-trust world, any workload touching sensitive data should fist prove where its runs, constantly! CanaryBit brings that principle to modern VM-based and Kubernetes environments.
