Cloud Native Security Whitepaper – Italian translation

Executive Summary

The industry has shifted towards cloud-native development and distribution models, leading to an expanding ecosystem of technologies, products, standards, and solutions that challenge decision-makers to keep pace with complex projects. The role of the Chief Information Security Officer (CISO) is evolving, focusing on aligning security practices with business value propositions. This shift has introduced new challenges in cloud-native security due to the rapid development and distribution cycles. Traditional perimeter-based security models, relying on static identifiers like IP addresses, have become impractical in this context.

The paper advocates for a paradigm shift in security strategies from static approaches to dynamic ones based on workload identification through attributes and metadata such as labels and tags. This enables the protection of cloud-native application workloads while adapting to the continuous flow of operations. The paper highlights the need for increased automation of security controls throughout the application lifecycle and emphasizes the importance of adopting a zero-trust architecture model for secure implementation.

The cloud-native application lifecycle is segmented into distinct phases: "Development", "Distribution", "Deployment", and "Runtime". Security measures must be integrated into these phases rather than being relegated to the end of the lifecycle. The paper outlines how modern security practices focus on developing code that adheres to recommended design patterns and ensures the integrity of the development environment.

Security in the distribution phase requires robust methods to verify not only the integrity of the workload but also the processes involved in its creation and operation. This is particularly challenging given the use of consistent, functional, and coherent open-source software and third-party runtimes, along with their dependencies. Automated scanning of artifacts like container images is crucial to ensure security against vulnerabilities, malware, insecure development practices, and other potential risks.

In the deployment phase, security is continuously validated and applied in real-time to workload attributes. This includes verifying signed artifacts, compliance with container image policies, and binary authorization policies in staging and production environments. The paper underscores the importance of secure observability features within the workload itself to monitor logs and metrics with high confidence.

Finally, the runtime environment in a cloud-native setting encompasses various components with associated security issues, such as hardware, providers, operating systems, networks, storage, container runtime, and orchestration. The container runtime consists of different isolation implementations at various levels, including shared kernels, micro-VM sandboxes, and trusted execution environment sandboxes. Choosing a runtime that meets specific security requirements is essential. For instance, a multi-tenant environment might require a VM-based sandbox, while sensitive financial data processing might benefit from dedicated hardware or Confidential Containers for encryption or virtual machine execution.

Best practices ensure that only approved processes operate within namespaces, safeguarding the overall security posture of cloud-native applications.