How does a shared responsibility model work in multicloud?

When an organization moves data or applications to the public cloud, its security responsibilities do not fully shift to the cloud provider. Instead, cloud users and providers adopt a shared responsibility model. As a general rule, the cloud provider is responsible for securing the cloud infrastructure, including the network, servers, databases and storage. At the same time, the cloud user maintains security responsibilities for apps, data and other IT components that are hosted on the cloud platform. Cloud providers like Amazon Web Services describe this as a "shared responsibility model."

The shared responsibility model is even more pronounced in multicloud computing because the attack surface is larger.

For example, suppose a cloud user provisions a virtual machine instance. The cloud provider must secure the server, storage and other underlying resources. To do this, the provider might update firmware or ensure its data center facilities are physically secure. However, it's the users' responsibility to securely deploy and configure the applications or operating systems running on that cloud platform.

In a shared responsibility model, the user must also set firewall and network configurations, implement the correct identity and access management (IAM) posture and perform other tasks. If unencrypted data is stolen through an open network port without a firewall, because the user did not configure those resources, the cloud provider is not responsible.

The shared responsibility model is even more pronounced in multicloud computing because the attack surface is larger; IT teams must account for multiple cloud infrastructures and providers. As organizations adopt and integrate multiple clouds, it's even more critical to establish secure data storage, implement comprehensive IAM models and address application security flaws.

Deployment automation will help organizations provision appropriate instances and deploy well-tested and properly configured base images, such as Amazon Machine Images. Similarly, log systems, such as Amazon CloudWatch, should gain popularity due to application tracking or data access for immediate alerting or forensic analysis of malicious behaviors.