Cybersecurity 3.0: How Confidential Computing Is Reinventing Cloud Trust
With the rapid development of AI and cloud
technologies, the data processed by enterprises on the cloud—ranging from
customer information and business secrets to AI training data—has become a
prime target for hackers and insider threats. Although most enterprises encrypt
data at rest and in transit, a major security gap remains: insufficient
protection when data is being processed.
This issue is particularly prominent in the
European and American markets. According to IBM’s 2024 Data Breach Report,
Europe—under the strict GDPR (General Data Protection Regulation) requirements
for traceability and confidentiality in data processing—has seen an increase in
privacy risks within cloud computing. In the U.S., regulations such as the
Department of Defense’s CMMC (Cybersecurity Maturity Model Certification),
HIPAA (Health Insurance Portability and Accountability Act) in healthcare, and
various state-level data protection laws are forcing companies to face a tough
question: Can we still trust the cloud to protect our data once it’s there?
Data is most vulnerable during
processing—when it is decrypted. Traditional cybersecurity models mainly
protect data at rest and in transit. However, once the data enters memory and
is processed by the CPU for analysis, inference, or computation, it usually
must be decrypted. In other words, during the most critical and valuable stage
of computation, data is actually most exposed and prone to attack.
What is Confidential Computing and Why
Is It Gaining Attention?
Confidential computing is a hardware-based
data protection technology. Its core concept is to encapsulate both the
computation process and the data being processed within an encrypted and
isolated memory enclave. This area functions like a "digital
safe"—even while the data is being used and processed, it remains
inaccessible to any external programs, cloud platform operating systems, or
even cloud administrators.
This eliminates the need to fully trust the
cloud platform. Instead, enterprises can create verifiable boundaries of
trust based on confidential computing. Technology giants such as NVIDIA,
Intel, and AMD have all embraced this technology, and cloud service
providers including Microsoft, Google, and Amazon have launched
corresponding solutions. For instance, in April of this year, NVIDIA and
Google collaborated to launch Google AI models for enterprise use on Google
Cloud, protected by NVIDIA’s confidential computing technology.
Confidential Computing Offers Three
Strategic Values:
- Rebuild Cloud Trustworthiness: The
cloud is no longer a source of risk, but a controllable resource. Data
sovereignty remains with the enterprise, not the platform.
- Enable Cross-Border Data Collaboration and Innovation: Different departments, organizations, and even countries can
collaborate on AI training without disclosing raw data—fostering the data
economy.
- Meet Regulatory and Audit Requirements: Faced with legal risks such as GDPR and HIPAA, companies can
provide verifiable evidence of secure data processing and pass compliance
checks effectively.
From a crisis of trust to data autonomy,
confidential computing marks a turning point—it is a technology that enables verifiable
trust. It does not replace the cloud but makes the cloud trustworthy;
it does not replace regulations, but enables them to be verified.
In the global wave of digital
transformation, enterprises must gain control over the computing process
in order to unlock the full value of their data. Mastering confidential
computing is the key to actively building trust.
During the Cybersecurity 1.0 era,
protection relied on mainframes; in Cybersecurity 2.0, the focus shifted to the
cloud, with software-based defense mechanisms. Now, in the era of Cybersecurity
3.0, it is necessary to establish a hardware-based "Root of
Trust", giving cloud platforms a technically verifiable foundation for
trust.
For Taiwanese Enterprises:
Confidential computing empowers businesses
to move beyond mere “data leakage prevention” toward “trusted data
collaboration,” balancing operational flexibility and regulatory
compliance. The earlier companies deploy this technology, the better
positioned they will be to lead in the data economy, build commercial trust,
forge data collaboration alliances, and secure a place in the global industrial
value chain.
沒有留言:
張貼留言