ANZA引入了新的Alpenglow共识协议，旨在使Solana网络更快，更确定性

Alpenglow是一种新的共识协议，是由以索拉纳（Solana）为中心的研究与开发公司Anza提出的。它旨在使网络更快，更确定性和更安全。

Anza, a Solana-focused research and development firm, has unveiled a new consensus protocol named Alpenglow. This protocol aims to make the Solana network faster, more deterministic, and safer. Described as the biggest consensus redesign since Solana’s launch, Alpenglow is set to replace the existing TowerBFT (Byzantine Fault Tolerance) mechanism and Proof of History with a more easily understandable and efficient system.

以索拉纳（Solana）为重点的研究与开发公司Anza揭示了一项名为Alpenglow的新共识协议。该协议旨在使Solana网络更快，更确定性和更安全。被描述为自Solana推出以来最大的共识重新设计，Alpenglow将用更容易理解和高效的系统来替换现有的TowerBft（拜占庭容错）机制和历史证明。

Targeting Performance Gains Without Compromising Network Safety

目标性能提高而不损害网络安全

目标性能提高而不损害网络安全

In his X post, Yakovenko further stated that he got nearly everything wrong about consensus, except the important parts. Drawing attention to the design of Alpenglow, he highlighted two critical requirements:

Yakovenko在他的X帖子中进一步表示，除了重要的部分，他几乎对共识几乎所有错误。引起人们对Alpenglow的设计的关注，他强调了两个关键要求：

* Consensus should not interfere with block producers using 100% of the available bandwidth at all times.

*共识不应始终使用100％的可用带宽来干扰块生产商。

* Users must experience deterministic finality in a single round.

*用户必须在一轮中经历确定性的终结性。

According to Anza’s blog, Alpenglow drastically reduces transaction finality latency to a median of 100-150 ms. In contrast, TowerBFT takes roughly 12.8 seconds from block creation until finality. This marks a 100x improvement, rendering it competitive with centralized infrastructure.

根据Anza的博客，Alpenglow将交易最终延迟延迟大大减少到中位数为100-150毫秒。相比之下，TowerBft从Block Creation到最终结束大约需要12.8秒。这标志着100倍的改善，使其在集中基础设施中具有竞争力。

Alpenglow achieves this by splitting the faults into two categories: full Byzantine and down or unavailable. Based on this fault model, the protocol can guarantee safe finality within one or two rounds. If a fork accumulates 40% of the total stake in votes (20% Byzantine and 20% down), then the node can safely commit to it. The fork will be finalized either in one round with 80% of votes or in two rounds with 60%.

Alpenglow通过将故障分为两个类别来实现这一目标：完整的拜占庭和下降或不可用。基于此故障模型，该协议可以保证一两轮内的安全终端。如果叉子积累了投票总股份的40％（拜占庭式20％和20％下降），则节点可以安全地承诺。叉子将以80％的选票或两轮为60％在一轮中最终确定。

Rotor and Votor: Building for the Future

转子和选民：未来建设

转子和选民：未来建设

This approach allows for deterministic finality without slowing down block production. Instead of handling epochs full of unconfirmed blocks, Alpenglow reduces the active state to just two blocks. This simplifies the system for developers and validators, especially when different parts of the network aren't perfectly synchronized.

这种方法允许确定性终结性，而不会减慢块生产。 Alpenglow并没有处理充满未经证实的块的时代，而是将活动状态减少到只有两个块。这简化了开发人员和验证器的系统，尤其是当网络的不同部分未完全同步时。

Yakovenko also noted that the protocol has simplified and formalized Turbine, now renamed Rotor. Together with Votor, Alpenglow's voting component, these components lay the foundation for enhancing scalability features, such as support for multiple concurrent leaders and increased validator efficiency.

Yakovenko还指出，该协议已简化和形式化的涡轮机，现在更名为转子。这些组件与Alpenglow的投票组成部分一起，为增强可扩展性功能（例如支持多个并发领导者的支持和提高验证者效率）奠定了基础。

Although still a proposal, if approved, Alpenglow could be a major step in Solana's ongoing evolution.

尽管仍然是一项提议，如果获得批准，Alpenglow可能是Solana持续发展的重要一步。