What is Confidential Computing and How Does It Work?
What is Confidential Computing?
In essence, confidential computing is a technology that should improve data privacy and security. The hardware-based innovation works with encrypted computations of sensitive data and isolates it from host resources. As a result, data remains separate from an application and other systems during its processing.
Processing sensitive data has quite a few obstacles. One is that it is quite difficult to ensure privacy is maintained during this process. That is where confidential computing should offer a solution. Big Tech companies have been enthusiastically supporting the technology since its infancy. In recent months, Meta and Nvidia have also made the leap to confidential computing.
How Does Confidential Computing Work?
Traditional security measures usually protect data in transit over the network or while it is in storage, but not while it is being processed. However, the confidential computing method uses so-called “enclaves” to prevent security breaches during processing. In geography, an enclave refers to an area of one state that is completely enclosed by another state (think, for example, Vatican City within Italy). Similarly, a security enclave is a highly controlled area within a larger secure system.
Organizations deploy security enclaves to data storage to identify and secure extremely sensitive material in a specialized, purpose-built vault.
Confidential computing uses a method backed by real-time encryption in system memory. In this way, a trusted execution environment (TEE) is created within a CPU. Unauthorized third parties cannot access this, so sensitive data and code remain free from exposure while remaining available for task processing.
A TEE resides on a device’s main CPU, isolated from the primary operating system. It guarantees that data is kept, processed and protected securely.
Enthusiasm and Skepticism about Confidential Computing
As mentioned, many tech giants are embracing the technology. They often expand their portfolio with platforms that incorporate confidential computing. There are plenty of examples of this: Arm, Dell, AMD, HPE, IBM, Google, Microsoft, Intel, Meta, Amazon and Nvidia. Smaller players will most likely follow in their footsteps. On top of that, experts predict that the market value of confidential computing will reach $54 billion by 2026.
However, there is also a lot of skepticism surrounding this innovation. This is because of its rapid popularity and the fact that mega corporations are its figureheads. Many view confidential computing as a marketing gimmick by cloud providers to promote memory encryption, with the premise that it could be the move that drives an organization to the cloud.
Those who oppose confidential computing usually argue that the technology does not solve larger problems. This particularly concerns the shortcomings of memory encryption: it does not address system images, updates, divergent services, and other security issues.