In today’s digital age, data has become a core resource, generating up to 402 million terabytes of sensitive data globally every day. With the continuous improvement of individuals’ awareness of data privacy, the demand for private computing of data is becoming increasingly urgent. Currently, the satisfaction of this demand largely depends on Amazon Web Services (AWS), which, with approximately 30% of the global cloud computing market share, has become a commonly used cloud computing platform for developers. However, AWS’s centralized architecture has gradually exposed many problems, prompting the industry to start exploring more secure and more suitable solutions for emerging application scenarios.
The current situation of AWS: Advantages and limitations coexist
As the leader in the field of cloud computing, AWS provides developers with a rich array of tools covering computing, storage, and database services, which can almost meet all the computing needs of developers. Its huge market share and high usage rate highlight its important position in the industry – nearly 48% of software engineers or developers use AWS in various ways.
With the continuous expansion of its customer base, including large financial institutions, government agencies, and start-ups, which attach great importance to data security, AWS’s security is facing a severe test. To address this issue, AWS has launched AWS Nitro Enclaves, aiming to protect the data in use and improve the encryption system for static data and data in transmission by introducing a Trusted Execution Environment (TEE).
AWS Nitro Enclaves is based on hardware and provides users with an isolated execution environment. It builds a secure space within Amazon EC2 instances, cuts off interactive access, persistent storage, and external network connections, and isolates sensitive workloads from parent EC2 instances and their operators, as well as other software, significantly reducing the attack surface. However, it cannot be ignored that Nitro Enclaves was created and managed entirely relying on AWS’s EC2 service. From its birth to its termination, all its management links were controlled by AWS infrastructure, and the highly centralized architecture became its biggest drawback.
Furthermore, there are a series of specific problems with AWS Nitro Enclaves. In terms of security, it connects multiple Nitro cards to the CPU to run the TEE payload, making both the underlying CPU and Nitro cards potential vulnerability points. The memory encryption mechanism is not perfect. The external characteristics of Nitro cards make it difficult to achieve end-to-end encryption of memory data, increasing the risk of data being tampered with during memory access.
At the developer usage level, the creation and configuration processes are complex and require the use of tools such as Docker, AWS CLI and Nitro Enclaves CLI, with a relatively high operational threshold. The reliance of TEE on Nitro cards leads to unreliable code integrity proofs; The trust model of AWS for the management of developer and administrator permissions also brings the risk of internal personnel accessing or tampering with sensitive data.
In terms of application adaptability, AWS Nitro Enclaves provide insufficient support for decentralized applications and ecosystems. It lacks persistent storage, which limits some application scenarios such as AI agents; Key management does not meet the “zero trust” requirements of Web3. Security code upgrades are not supported. Developers need to build related functions by themselves, which is time-consuming and prone to vulnerabilities. Network access restrictions make it difficult to set up Web services, and developers need to write a large amount of code to ensure the security of the application.
The Web3 era calls for new solutions
The core concept of Web3 is to build a trustless system where users have absolute control over data and are cautious about developers and administrators. Against such a background, the traditional AWS Nitro Enclaves clearly cannot meet the requirements. Web3 participants expect their applications to be secure at the hardware level, ensuring that data is not interfered with, crawled or altered during memory access and other links. At the same time, they require clear guarantee and verification mechanisms for data operations.
Decentralized TEE Cloud: Innovative Exploration of Phala Network
To fill this gap, Phala Network proposed the concept of decentralized TEE cloud. Decentralized cloud stores, processes and manages data in a network composed of multiple nodes, operates relying on blockchain technology, and discards the single control entity model of traditional centralized cloud systems. On this basis, the decentralized TEE cloud combines TEE with a decentralized node network. Each node is equipped with a TEE to ensure that sensitive code and data remain confidential and intact even when facing node operators.
Phala Network has constructed a network composed of decentralized worker nodes, which perform computing tasks based on user requirements, such as running smart contracts and processing sensitive data. After users deploy applications or tasks to the network, the computing process is completed within the TEE of the node. The correctness of the computing is verified through cryptographic proofs, and economic rewards are given to encourage node operators to provide honest and secure services to ensure the stable operation of the network.
Compared with AWS, Phala Cloud shows significant advantages in multiple aspects. In terms of deployment convenience, the setup process of AWS Nitro Enclaves is cumbersome, involving time-consuming operations such as installing multiple tools and converting image files. Phala supports “migrate and modify”. With the help of the Dstack SDK, developers only need to make a few modifications to the existing Docker containers to convert them into confidential virtual machines and complete the deployment through the friendly Cloud UI. At the same time, it is compatible with templates and custom Docker Compose files.
In terms of security, AWS relies on users’ trust in the configuration of developers and administrators. The centralized architecture poses a risk of data being accessed by internal personnel. Phala adopts a zero-trust model to fundamentally prevent unauthorized access to sensitive data by all parties, which is more in line with the needs of Web3 applications.
In terms of product positioning, AWS mainly serves enterprise customers, and its technical and product features deviate from the demands of Web3 start-ups. Phala is specifically designed for decentralized applications, natively supporting the interaction between AI agents and blockchain smart contracts, as well as privacy-protected Dapps. It is deeply integrated into the blockchain ecosystem and has established partnerships with numerous protocols.
Compared with other TEE solutions, Phala Network is also unique. Oasis Protocol and Secret Network focus on implementing privacy smart contracts using TEE in their own blockchains, while Phala provides a decentralized cloud computing platform for offline computing across networks. The Marlin Protocol focuses on enhancing the performance of Web3 networks and does not involve computing and privacy functions. Phala, with its extensive TEE hardware support, such as Intel SGX, Intel TDX, AMD SEV and NVIDIA GPU TEE, as well as the developer-centered tool Dstack, has become the only fully decentralized TEE cloud platform.
In addition, Phala innovatively integrates TEE with multi-party Computation (MPC) to build a decentralized Root of Trust (DeRoT) model. By running MPC within TEE, it reduces the risk of node collusion. Combined with multiple proof methods, it further enhances the security of TE-based applications.
Future outlook: Broad application scenarios
Decentralized TEE cloud will have a rich array of application scenarios in the future. In the field of artificial intelligence, it is supported to deploy AI models within the TEE to ensure their security and autonomy when interacting with the blockchain network. Developers can achieve truly cension-resistant and privacy-protecting AI computing with the help of GPU TEE. In terms of the migration of traditional applications, they can be placed in a secure and trustless environment to enhance security. In the field of data analysis, privacy-protected data analysis can be achieved. In decentralized finance (DeFi), enhance secure computing capabilities, such as keeping trading positions confidential and supporting dark pool transactions; It is also possible to achieve MEV (Maximum Extractable Value) protection by moving block construction into TEE, ensuring fair sequencing and execution of transactions.
From the limitations of AWS to the rise of decentralized TEE cloud, the cloud computing field is undergoing a profound transformation. The emergence of innovative solutions such as Phala Network has provided new possibilities for data security and application development in the Web3 era. With the continuous advancement of technology and the in-depth expansion of applications, decentralized TEE cloud is expected to reshape the landscape of the cloud computing industry, providing users with more secure, reliable, and emerging demand-adapted computing services.
Related Topics:
