Jump to Content
Security & Identity

Better together: Expanding the Confidential Computing ecosystem

December 16, 2020
https://storage.googleapis.com/gweb-cloudblog-publish/images/Security_BlogHeader_B_epmyJP1.max-2600x2600_IkKZwmx.jpg
Nelly Porter

Director, Product Management, Google Cloud

Sam Lugani

Group Product Manager, Google

As of June 16, 2022, Confidential VMs are generally available on general purpose N2D machine types and compute optimized C2D VM families. Learn more here.


Core to our goal of delivering security innovation is the ability to offer powerful features as part of our cloud infrastructure that are easy for customers to implement and use. Confidential computing can provide a flexible, isolated, hardware-based trusted execution environment, allowing adopters to protect their data and sensitive code against malicious access and memory snooping while data is in use. 

Today, we are happy to announce that we have completed the rollout of Confidential VMs to general availability in nine regions. Our partners have played a huge part in this journey. They have been critical in establishing an ecosystem that aims to make Confidential Computing ubiquitous across mobile, edge, and cloud. We spoke to Raghu Nambiar from AMD, Mark Shuttleworth from Canonical, Burzin Patel from HashiCorp, Mike Bursell from Red Hat, Dr. Thomas Di Giacomo from SUSE, and Solomon Cates from Thales. Here are the excerpts.

https://storage.googleapis.com/gweb-cloudblog-publish/images/amd1.max-600x600.jpg

Raghu Nambiar, Corporate Vice President, Data Center Ecosystems, AMD

Confidential Computing is a relatively new concept with a goal to encrypt data in use in the main memory of the system, while still offering high performance. Confidential Computing addresses key security concerns many organizations have today in migrating their sensitive applications to the cloud and safeguarding their most valuable information while in-use by their applications. It wouldn’t be a surprise if in a few years, all virtual machines (VMs) in the cloud are Confidential VMs. 

How did you approach confidential computing?
The 2nd Gen AMD EPYC processors used by Google for its Confidential VMs uses an advanced security feature called Secure Encrypted Virtualization (SEV). SEV is available on all AMD EPYC processors and, when enabled by an OEM or cloud provider, it encrypts the data-in-use on a virtual machine, helping to keep it isolated from other guests, the hypervisor and even the system administrators. The SEV feature works by providing each virtual machine with an encryption key that isolates guests and the hypervisor from one another, and these keys are created, distributed, and managed by the AMD Secure Processor. The benefit of SEV is that customers don’t have to re-write or re-compile applications to access these security features 

With SEV-enabled Confidential VMs, customers have better control of their data, enabling them to better secure their workloads and collaborate in the cloud with confidence. 

What kind of performance can we expect?
What’s really impressive about Google Confidential VMs powered by AMD EPYC processors with SEV enabled is that it offers performance close to that of non-confidential VMs. AMD and Google’s engineering teams ran a set of well-known application benchmarks for relational database, graph database, webserver as well as Computational Fluid Dynamics and popular simulation workloads in FSI on Google Confidential VMs and Google’s N2D VMs, of which Confidential VMs are based on. The difference in using SEV versus not using SEV on the applications listed, was measured to be just a small overhead in application performance. 

Any final thoughts?
Confidential Computing is a game-changer for computing in the public cloud as it addresses important security concerns many organizations have about migrating their sensitive applications to the cloud. Google Confidential VMs, with AMD EPYC processors and SEV, strengthen VM isolation and data-in-use protection helping customers safeguard their most valuable information while in-use by applications in the public cloud. This is a paradigm shift and we’re excited to work with Google to make this possible.

https://storage.googleapis.com/gweb-cloudblog-publish/images/canonical.max-600x600.jpg

Mark Shuttleworth, CEO, Canonical

Confidential Computing directly addresses the question of trust between cloud providers and their customers, with guarantees of data security for guest machines enforced by the underlying hardware of the cloud. With Google’s addition of Confidential Computing to multiple regions, customers gain a secure substrate for large-scale computation with sensitive data and a path to regulatory compliance for new classes of workload on the cloud.

What value does the partnership between GCP and Canonical create?
Close technical collaboration between Google and Canonical ensures that Ubuntu is optimized for GCP operations at scale. Confidential Computing requires multiple pieces to align and we are delighted to offer full Ubuntu support for this crucial capability at the outset with Google.

How will this benefit organizations?
Organizations gain peace of mind that large classes of attack on cloud guests are mitigated by Confidential Computing. Memory encryption with hardware key management and attestation prevents a compromise of the hypervisor becoming a compromise of guest data or integrity. Customers can now consider GCP as secure as private infrastructure for a much wider class of workloads. Canonical Ubuntu fully supports Confidential Computing on Google Cloud, providing a new level of trust in public cloud infrastructure.

https://storage.googleapis.com/gweb-cloudblog-publish/images/hashiCorp.max-600x600.jpg

Burzin Patel, Vice President of Global Alliances, HashiCorp

HashiCorp Vault enables teams to securely store and tightly control access to tokens, passwords, certificates, and encryption keys for protecting machines and applications. When combined with GCP’s Confidential Computing capabilities, confidentiality can be extended to the HashiCorp Vault server’s system memory, ensuring that malware, malicious privileged users, or zero days on the host cannot compromise data. 

Why did you choose Google Cloud as a partner for Confidential Computing?
Google Cloud’s Confidential Computing nodes operate exactly like regular compute nodes making the offering very easy to use. We were able to take our existing Vault binary and host it on the Confidential Computing node to leverage the confidential computing benefits. No code or configuration changes were needed.

What is the gap confidential computing solves specifically for your customers?
Vault stores all of its sensitive data in memory and is stored as plaintext. In the past there were no easy solutions to keep this runtime memory protected. However, with the availability of confidential computing nodes, the data in memory is protected via encryption by utilizing the security features of modern CPUs together with confidential computing services.

Any use-cases that are top of mind for you when it comes to confidential computing?
HashiCorp Vault allows organizations to eliminate system complexity where any mistakes or misconfiguration could lead to a breach or data-leakage that in turn can halt operations and erode trust across customers. Together, HashiCorp Vault and Google Cloud’s Confidential Computing help organizations manage their most critical secrets and assets. This includes the entire secret lifecycle, from the initial creation, to sharing and distribution, and to the revocation or expiration of credentials and secrets.

Any final thoughts?
Security is the most critical element for enterprise customers looking to adopt the cloud. Customers are looking for a flexible solution that is robust and highly secure. The combination of HashiCorp Vault and Google Cloud Confidential Computing provide users a critical solution for their enterprise-wide cloud security needs.

https://storage.googleapis.com/gweb-cloudblog-publish/images/red_hat.max-600x600.jpg

Mike Bursell, Chief Security Architect, Red Hat

As more businesses and organizations move to the cloud, security remains a top priority. Maintaining the same levels of confidentiality that their partners, customers, regulators and shareholders expect across private and public clouds is vital. Red Hat believes that Confidential Computing is one key approach to extend security from on-premises deployments into the cloud, and Google’s announcement of Confidential VMs is an example of how customers can further secure their applications and workloads.

What has Red Hat’s approach been to Confidential Computing?
Red Hat Enterprise Linux is an enterprise operating system designed to handle the needs of customers across on-premises and hybrid cloud environments. Customers need stability, predictability and management solutions that scale with their workloads, which is why we enable Confidential Computing solutions in our product portfolio. That way customers don’t have to worry about migration costs. 

How will confidential computing impact cloud adoption?
Often, customers with regulatory concerns have greater concerns about shifting into a truly open hybrid cloud environment, as they cannot expose their more sensitive data and applications outside their own data centers. Red Hat believes Confidential Computing can help them make this shift, expanding their opportunities for digital transformation, allowing them to provide quicker, more scalable and more competitive solutions, while maintaining the data privacy and protection assurances that their customers expect and require. As organizations balance the need for security with the opportunities presented by the cloud, Confidential Computing provides new ways to safely and securely embrace those opportunities.

https://storage.googleapis.com/gweb-cloudblog-publish/images/suse.max-600x600.jpg

Dr. Thomas Di Giacomo, Chief Technology & Product Officer, SUSE

Confidential VMs is a cloud industry security game-changer. This offering for our joint cloud customers expands sensitive data protection and compliance requirements, especially for regulated industries. The best part is you can run legacy and cloud-native workloads securely without any refactoring to the underlying application code, simplifying the transition to the cloud, all with little to no performance penalty.

How has SUSE been working with Google Cloud and AMD?
Working closely with AMD, SUSE added upstream support for AMD EPYC SEV processor to the Linux Kernel and was the first to announce Confidential VM support in SUSE Linux Enterprise Server 15 SP1 available in the Google Cloud Marketplace. These innovations allow our customers to take advantage of the scale and cost savings of Google Cloud Platform and the mission-critical manageability, compliance, and support from the #1 rated Linux support team, SUSE.

How do you foresee this benefiting organizations?
Confidential VMs will help tremendously accelerate our customer migrations to the cloud on their hybrid cloud digital transformation journey. This technology opens up new areas of migration opportunities for legacy on-premises workloads, custom applications as well as Private and Government workloads that require the utmost security and compliance requirements once considered not cloud-ready in the past.

https://storage.googleapis.com/gweb-cloudblog-publish/images/thales.max-600x600.jpg

Solomon Cates, Principal Technologist, CTO Office, Thales

Confidential computing is a fundamental step in providing users control of their data as it goes “off premise” into cloud environments and all the way to the edge. Customers can essentially transition their workloads to the cloud with high assurance that includes auditable “proof” of control. And, architecturally, it opens up so many possibilities for customers.

Many enterprises have significant trepidation when it comes to security in the cloud. Confidential computing helps alleviate that. For example, security professionals no longer have to worry about a cloud provider seeing or using their data.

How does Confidential Computing help your customers?
Confidential computing solves an issue that enterprises specifically have around trust in memory—namely that memory cannot be seen or used by a cloud provider. Three key use cases that can immediately benefit from this technology include edge computing, external key management and in-memory secrets.

What made you partner with Google Cloud?
Thales and Google Cloud have collaborated across a number of areas including cloud, security, Kubernetes containers and new technologies such as Continuous Access Evaluation Protocol (CAEP).  At the core, we both strive to offer customers the best option for strong security and privacy protection.

Any final thoughts?
From both a strategic and technical standpoint, Thales and Google Cloud have a shared vision that focuses on customer control and security of their data in the cloud. Through our work around confidential computing, we will bring new possibilities for securing workloads at the edge. Together, we are making it possible for enterprises to put their trust in the cloud with more sovereign control over their data security.

We thank our hardware and software partners for their continuous innovation in this space. Confidential Computing can help organizations ensure the confidentiality of sensitive, business critical information and workloads, and we are excited to see the possibilities this technology will open up for your organization.

Posted in