Some of the biggest barriers to cloud adoption are security concerns: data loss or leakage, and the associated legal and regulatory concerns with storing and processing data off-premises.
In the last 18 months, 79% of companies have experienced at least one cloud data breach; even more alarmingly, 43% have reported 10 or more breaches in that time. Despite the clear advantages of cloud infrastructure, one of the main challenges that often gets overlooked is the need to: (1) trust that the infrastructure will be secure enough against threats and (2) that the chosen cloud provider wonât purposefully or inadvertently access the data processing on their infrastructure. When dealing with highly sensitive/confidential data (such as banking information or healthcare patient data), this becomes a major concern and a barrier to further cloud adoption.
Traditional approaches for protecting data have relied upon implementing access controls and policies and encrypting data at rest and in transit, but none are able to prevent the threat in its entirety because a fundamental challenge remains: keeping data encrypted when in use, while it is being processed. Confidential computing â projected to be a $54B market by 2026 â is emerging as a way to remove the need for trusting infrastructure and service providers by keeping data protected/encrypted even when in use.
Confidential computing technology uses hardware-based techniques to create isolated environments called enclaves (also known as Trusted Execution Environments or TEEs).
Code and data within enclaves are inaccessible by other applications, users, or processes colocated on the system. The enclave keeps the data encrypted even when in use â while in memory and during computation. With a secure enclave environment, multiple parties can collaborate on analytics and AI use cases without compromising the confidentiality of their individual data and exposing it to other parties.
According to a recent survey, using secure enclaves in the enterprise setting is attractive for implementing safeguards for the following scenarios:
- Protect against insider threats. Data in the cloud is accessible to the database administrators of the cloud applications or infrastructure via direct access to the database, application logs, and device memory
- Prevent platform software (i.e., a platform hypervisor) from accessing data
- Protect data from adjacent workloads in a multitenant/user environment
- Protect the integrity of crowdsourced ML models
- Confidential data sharing and multi-party collaboration
If these scenarios apply to you and your business, but youâre unsure what youâll need to know to get started, here are five questions to ask your CISO:
1. Will I need to deploy specialized hardware to keep our data protected?
Confidential computing technology is now available on all major cloud providers. This obviates the need to procure and maintain specialized hardware yourselves. Even though confidential computing and secure enclaves are still in the âemerging technology bucket,â organizations can easily adopt confidential computing through cloud vendors and ISVs. The cloud providers see the benefit of secure enclaves and their future potential as a transformative technology, and so have bought in.
2. Will we need to rewrite applications to use secure enclaves?
Some confidential computing technologies, such as Intel SGX, require application modifications before they can run within enclaves. Other technologies, such as Confidential VMs, provide more flexibility and can run unmodified applications.
But, from a security perspective, this has the downside of having to trust the entire software stack within the VM. So, depending on the use case and requirements, one technology may be preferable over the other. In addition, proper adoption of confidential computing requires orchestrating management of the other constituent technologies, such as remote attestation.
The enclave adoption process can be complex and engineering teams will have to take time to build these capabilities to get their applications up and running. While bandwidth may be tight at times, the ROI is worth it in the long run. A growing ISV ecosystem can also help in the seamless adoption of confidential computing for a rich variety of use cases.
3. Can I use secure enclaves to improve data collaboration with other teams?
Before data can be shared with other teams, organizations typically need to follow a cumbersome governance process to restrict access to sensitive data, eliminate data sets or mask specific data fields, and prevent any level of data sharing.
Integrating secure enclaves provides an opportunity for organizations to increase both productivity and security measures. Multiple data owners can individually encrypt their entire data (including PII), pool it together, and analyze the collective data set within enclaves. Done effectively, multi-party collaboration can drive faster business results by enabling new and higher-quality insights.
4. Will I need to add additional security expertise to the team?
Implementing confidential computing workflows can be difficult to do directly without using existing tools and software. One needs to make sure that confidential data is protected throughout its lifecycle. This can have a variety of moving parts â from integrating with existing key management systems to managing secure enclave infrastructure, rewriting applications, deploying code securely and verifiably to the enclaves, and keeping confidential data encrypted in storage and in transit in/out of the enclaves. However, there is a rich emerging ISV ecosystem of software that alleviates the complexities of confidential computing for a rich variety of use cases, making it easy to use and adopt by non-experts.
5. Will I need to lock myself into a single cloud?
The top CPU vendors all introduced secure enclave and confidential computing solutions in recent years. These were adopted by the leading cloud vendors, some of which now offer solutions based on the same underlying technology. Microsoft Azure and Google Cloud Platform, for example, offer solutions based on AMDâs SEV technology. As software solutions running on top of these cloud platforms evolve, application vendors will introduce cross-platform solutions powered by the common hardware layers.
Conclusion
Businesses considering adopting cloud technology can better do so with secure enclaves. By asking your CISO these five questions, businesses can move into the future, understand what implementing secure enclaves will look like, better secure their data, and create a more efficient analytics process.
This ongoing shift to the cloud will increase efficiency for companies and reduce human error â especially knowing 57% of businesses will move their workloads to the cloud before the end of the year. When secure enclaves are implemented properly, the crucial component of ensuring security is not sacrificed. All businesses working with data should consider integrating confidential computing into their models to allow for analytics and AI on encrypted data.
Secure Processors Part I: Background, Taxonomy for Secure Enclaves and Intel SGX Architecture