
Briefing
The core research problem Fides addresses is the inherent trade-offs in existing asynchronous Byzantine Fault Tolerant (BFT) consensus protocols, particularly concerning scalability, communication overhead, and censorship resilience. Fides proposes a foundational breakthrough by integrating Trusted Execution Environments (TEEs) to offload critical consensus components, thereby achieving linear message complexity, guaranteed censorship resistance, and a reduced quorum size. This new theory fundamentally redefines the pathway to building highly performant and robust blockchain architectures, offering a blueprint for systems that can sustain high transaction throughput without compromising security or decentralization.

Context
Before Fides, asynchronous BFT consensus protocols grappled with significant limitations. These protocols typically required large quorum sizes, often three times the number of Byzantine replicas, to maintain security. They also suffered from high communication costs and relied on computationally expensive cryptographic primitives, such as global common coins, to reach agreement in asynchronous network conditions.
Furthermore, many existing designs exhibited poor censorship resilience, jeopardizing the liveness guarantee essential for continuous operation in adversarial environments. These challenges collectively hindered the practical deployment of truly scalable and robust decentralized systems.

Analysis
Fides introduces a novel asynchronous DAG-based BFT consensus protocol that fundamentally rethinks how distributed agreement is achieved. Its core mechanism involves abstracting and offloading four critical components ∞ Reliable Broadcast, Vertex Validation, Common Coin, and Transaction Disclosure ∞ into Trusted Execution Environments (TEEs). By confining these sensitive operations within TEEs, Fides dramatically reduces the Trusted Computing Base (TCB) compared to porting the entire protocol. This architectural innovation allows Fides to achieve linear message complexity, a significant improvement over quadratic or cubic complexities in prior BFT protocols.
The use of TEEs also guarantees censorship resilience by ensuring that transactions are processed fairly, and enables a 2x larger quorum size for fault tolerance while using a lightweight common coin. This approach fundamentally differs from previous methods by leveraging hardware-level security to circumvent traditional cryptographic and communication bottlenecks, thereby enhancing both efficiency and security simultaneously.

Parameters
- Core Concept ∞ DAG-based BFT Consensus
- New System/Protocol ∞ Fides
- Key Authors ∞ Shaokang Xie et al.
- Key Technology ∞ Trusted Execution Environments (TEEs)
- Message Complexity ∞ Linear
- Quorum Size Improvement ∞ 2x larger
- Throughput (Geo-distributed) ∞ 400k transactions per second
- Throughput (Local) ∞ 810k transactions per second

Outlook
The integration of Trusted Execution Environments within core consensus protocols, as demonstrated by Fides, opens significant new avenues for research and real-world applications. Future work will likely explore the broader applicability of TEEs in other distributed system primitives, optimizing their performance, and formally verifying their security guarantees within more complex blockchain architectures. In the next 3-5 years, this theory could unlock the development of highly scalable and censorship-resistant public blockchains capable of supporting mainstream adoption. It also paves the way for new research into hybrid consensus models that combine hardware-assisted security with cryptographic proofs to achieve unprecedented levels of performance and resilience in decentralized networks.