DA层的崭新时代 数据可用性层的技术突破与应用
作者：Bridget Harris 来源：substack 翻译：善欧巴，比特币买卖交易网
数据可用性层已成为模块化架构的重要组成部分，充当可插入组件来降低成本并扩展区块链。DA层的核心功能是确保链上数据可供所有网络参与者使用和访问。从历史上看，每个节点都必须下载所有交易数据以验证数据是否可用——这是一项效率极低且成本高昂的任务。这就是大多数区块链目前的工作方式，并且是可扩展性的障碍，因为验证所需的数据量随着块大小线性增加。最终用户在这里遭受损失：数据可用性成本占用户在 Rollup 上进行交易所产生的交易成本的90% （90%目前 Rollup 将交易数据发送到以太坊的成本为 1300-1600 美元/MB ）。
数据可用性采样 (DAS) 的引入从根本上改变了这种架构。通过 DAS，轻节点可以通过参与区块数据的多轮随机采样来确认数据可用，而不必下载每个整个区块。一旦完成多轮采样，并且达到数据可用的某个置信阈值，交易过程的其余部分就可以安全地进行。通过这种方式，链可以扩展其块大小，同时保持简单的数据可用性验证。并且还实现了可观的成本节省：这些新兴层可以将DA 成本降低高达 99% reduce DA costs by up to 99%。
除了实现更高的吞吐量之外，数据可用性层对于提高互操作性也很有意义。廉价的 DA 将不可避免地推动新的定制 Rollup 链的寒武纪爆发，通过 Caldera、AltLayer 和 Conduit 等 Rollup-as-a-service 提供商使部署变得越来越简单。然而，随着 L2 和 L3 生态系统的出现，它们默认会变得支离破碎。让用户使用新平台已经很困难了——如果互操作性、流动性和网络效应有限，情况会变得更糟。统一的 DA 层作为每个网络的基础，会使得资金流变得更加简单，并吸引更广泛的用户群。
CalderaCaldera和其他 RaaS 提供商将使项目能够在构建自定义Rollup时选择 DA 层
Avail Avail、, EigenDAEigenDA和, and CelestiaCelestia是 DA 生态系统中的主角——每个都服务于相同的空间，但在基础设施堆栈、执行和上市方面采取的方法略有不同。
在技术架构方面，Avail、Ethereum和EigenDA使用KZG承诺，而Celestia则使用欺诈证明来确认区块编码正确to confirm blocks are encoded correctly。生成 KZG 证明——虽然是证明 DA 的一种非常严格的方法——会给区块生产者带来更多的计算开销，特别是当区块大小增加时。另一方面，Celestia 假设数据可以通过其防欺诈方案隐式获得。作为没有完成计算“工作”的交换，系统必须等待一段时间来进行欺诈证明争议期，然后节点才能确认该块已准确编码。KZG 证明和欺诈证明都在经历快速的技术进步；它们的权衡可能会继续变得更加复杂，目前还不清楚一种机制是否会严格优于另一种机制。
对于 Avail 来说，他们具有 KZG 承诺的架构使他们非常适合 zk 结构——这是 Celestia 可能面临困难的领域，因为如果 zk 在未来占主导地位，他们依赖乐观证明。此外，Avail 的轻客户端 p2p 网络即使在所有全节点都宕机的情况下也可以支持该网络；而在 Celestia 的架构中，轻客户端在没有完整节点的情况下无法运行。Avail 和 Celestia 在 DAS（分布式存储）下均采用纠删码，将数据分割成碎片，增加冗余，并允许对数据进行重建以进行验证。
与 Celestia 和 Avail 的堆栈相比，EigenDA 利用了以太坊现有的基础设施。EigenDA 如果数据需要发送到汇总合约以证明数据可用，EigenDA继承与以太坊相同的最终确定时间。inherits但是，如果 rollup 完全使用 EigenLayer，则可以更快地实现最终确定。
为了达成共识，Avail 使用从 Polkadot SDK 继承的For consensus, Avail uses BABE + GRANDPABABE + GRANDPA以及指定的权益证明 (NPoS)。NPoS 用于提名serves to nominate委托人愿意选举的一组验证者，而 BABE 则规定谁将提议下一个区块，而 GRANDPA 则充当区块最终确定算法。
Celestia 使用 Tendermint 作为共识机制，允许用户质押他们的 TIA 以获得验证者质押奖励。尽管 Celestia 能够通过 Tendermint 实现快速确定性，但由于其 optimistic 架构，实际数据可用性的保证存在等待期（用户必须有时间提交欺诈证明）。
EigenDA本身没有共识，而是有两种机制来保证数据可用性的有效性：
监护权证明。这本质上是一种经济安全机制，可确保节点存储数据，但实际上并不能保证该数据提供给网络中的每个人。如果节点不遵守规定，例如，如果它们无法证明自己拥有数据，则节点将被削减。
充分的权力下放。确保运营商集保持去中心化和防串通对于网络的正确运行至关重要。有了庞大且独立的验证人集，数据服务就成为许多市场参与者愿意参与的竞争。在这种规模下，共谋是极其困难的。
值得一提的一个有趣观点是，Celestia 的活跃验证者集由前 100 名质押令牌最多的验证者组成，而这个门槛可能在未来降低。此外，他们的每个验证者都存储了整个数据集。而 EigenDA 将对存储一小部分数据的每个节点（未来可能是数百万个）进行优化，因此，如果足够的节点是诚实的，数据就可以被重建。有关 EigenDA 的完整起源（和更多细节）可以在 Sreeram 最近的帖子中找到。
最后，Avail 对主要 DA 层的核心组件进行了有益的比较。
关于每种设计的权衡也出现了新的讨论。David Hoffman 指出，Celestia 本身就是一个完整的区块链——一个复杂的堆栈，需要的不仅仅是纯粹的 DA。另一方面，EigenDA 只是一组智能合约，但它依赖于以太坊，而 Celestia 和 Avail 则不然。
Celestia 团队认为，为了安全性，代币是必要的，而 EigenDA 最终将需要代币，因为不可能削减链上数据的可用性。他们认为，为了确保节点是诚实的、数据可用并惩罚恶意节点，网络必须能够通过包括原生代币在内的激励结构进行验证。在这里，Celestia 的 Nick White 提出了对 EigenDA 的批评：「除非源链被分叉，否则保留数据的重新验证器不会被 slash——这是极不可能的，因为这是以太坊。」
从品牌角度来看，EigenDA 是一款与以太坊极其一致的产品。EigenLayer 团队正在根据 EIP-4844 和 danksharding 进行构建——用 Sreeram 的话说，EigenDA 被构建为「唯一以 ETH 为中心的数据可用性层」。他解释说，根据定义，数据可用性层是一种模块化产品，但其他 DA「Layer」实际上是区块链本身。
将 DA 层打包到区块链中确实会给在其上本地运行的 Rollups 带来明显的好处，主要是安全保证的形式。然而，Sreeram 提到，他的团队构建 EigenDA 的目标是创建一种产品，从第一原则开始为以太坊生态系统提供数据可用性服务——一个毗邻以太坊生态系统的真正的「Layer」。他指出，这里不需要单独的共识，因为基于以太坊的 Rollup 已经依赖网络进行排序和共识。（Sreeram 在最近的 Bankless 节目中解释了这一点。）
Avail 采用有效性证明和 DAS 构建，可在生态系统方面实现高度的灵活性和互操作性。他们的架构为可扩展框架奠定了基础，该框架旨在支持跨许多不同平台的服务。这种「不偏不倚」的立场允许更大的互操作性和资金流动，并且也吸引非以以太坊为中心的生态系统。这里的最终目标是从所有链中获取有序交易数据，并将其聚合到 Avail，使它们成为所有 web3 的协调中心。为了启动该网络，Avail 最近在其激励测试网的同时推出了节点冲突活动，允许用户运行验证器和轻客户端并参与网络挑战。
Celestia 的生态系统由 RaaS 提供商、共享排序器、跨链基础设施等组成，涵盖以太坊、以太坊 rollups、Cosmos 和 Osmosis 等生态系统。
Celestia 生态系统页面的快照
这些设计选择中的每一个，无论是技术方面还是营销方面，都伴随着有趣的权衡。就我个人而言，我不确定数据可用性类别是否会是一个赢家通吃或商品化的市场——相反，可能存在寡头垄断式市场，项目会选择最适合其需求的 DA 层。根据协议的类型，团队可以针对互操作性、安全性或对某个生态系统或社区的偏好进行优化。如果自定义用例汇总按预期爆炸式增长，他们会毫不犹豫地集成 DA 层 - 并且将有不止一种强大的选项可供选择。
这项技术——以及总体上的模块化叙述——仍然相对较新，Celestia 最近刚刚上线，Avail 和 EigenDA 将在未来几个月内进入主网。然而，迄今为止模块化主义的技术进步非常出色（其中许多概念在几年前还只是想法！）。通过从本质上改进我们构建和使用区块链的方式，DA 层无疑将成为本周期及以后周期的核心技术之一。
The data availability layer of Shanouba Bitcoin trading network has become an important part of modular architecture, which acts as an insertable component to reduce costs and expand the blockchain layer. The core function of the blockchain layer is to ensure that the data on the chain can be used and accessed by all network participants. Historically, every node has to download all transaction data to verify whether the data is available, which is an extremely inefficient and expensive task. This is the current working mode of most blockchains and is scalable. The obstacle is that the amount of data required for verification increases linearly with the block size, and the end user suffers losses here. The cost of data availability accounts for the transaction cost generated by users conducting transactions on the Internet. At present, the cost of sending transaction data to Ethereum is US dollars. The introduction of data availability sampling has fundamentally changed this architecture. Through light nodes, data can be confirmed to be available by participating in multiple rounds of random sampling of block data without downloading each whole block. Once multiple rounds of sampling are completed and the number is reached, According to an available confidence threshold, the rest of the transaction process can be safely carried out. In this way, the chain can expand its block size while maintaining simple data availability verification, and considerable cost savings have been achieved. These emerging layers can reduce the cost as high as possible. In addition to achieving higher throughput, the data availability layer is also very meaningful for improving interoperability. Inevitably, the Cambrian explosion of new customized chains will be promoted inexpensively, and the deployment will become more and more obvious through waiting for providers. Simple, however, with the emergence of the ecosystem, they will become fragmented by default, making it difficult for users to use the new platform. If the interoperability, liquidity and network effect are limited, the situation will become worse. As the basis of each network, a unified layer will make the capital flow simpler and attract a wider user base and other providers, which will enable the project to choose the layer and the protagonist in the ecosystem when building customization, each serving the same space, but executing in the infrastructure stack. The method adopted in listing is slightly different in terms of technical architecture and use commitment, while fraud proof is used to confirm that the block code is correctly generated. Although it is a very strict method of proof, it will bring more calculation expenses to the block producers, especially when the block size increases. On the other hand, it is assumed that the data can be implicitly obtained through its anti-fraud scheme. As an exchange system that has not completed the calculation work, it must wait for a period of time for fraud proof dispute period before the node can confirm. Both the accurate coding certificate and the fraud certificate are experiencing rapid technological progress, and their trade-offs may continue to become more complicated. It is not clear whether one mechanism will be strictly superior to the other. For me, their committed architecture makes them very suitable for the structure, which is an area that may face difficulties, because if they dominate in the future, they rely on optimism to prove that other light client networks can support the network even when all nodes are down. In the architecture, the light client can't run without complete nodes, and erasure codes is used to divide the data into pieces to increase redundancy and allow the data to be reconstructed for verification. Compared with the stack of sum, the existing infrastructure of Ethereum is used. If the data needs to be sent to the summary contract to prove that the data can be inherited with the same final determination time as Ethereum, but if it is fully used, the final determination can be realized faster. In order to reach a consensus, the inherited from is used. And the specified certificate of rights is used to nominate a group of verifiers that the client is willing to vote for, while it stipulates who will propose the next block and acts as the final determination algorithm of the block. The algorithm is used as a consensus mechanism to allow users to pledge theirs to obtain the verifier pledge reward. Although it can achieve rapid certainty, users must have time to submit fraud certificates because there is a waiting period to ensure the actual data availability of their architecture. There is no consensus in itself, but there are two mechanisms to ensure the effective monitoring of data availability. Authentication is an economic security mechanism in essence, which can ensure that nodes store data, but it can't actually guarantee that the data will be provided to everyone in the network. If nodes don't comply with the regulations, for example, if they can't prove that they own data, they will be cut off. It is very important for the correct operation of the network to ensure that the operator set remains decentralized and collusion prevention. With a large and independent verifier set, data services have become a competition that many market participants are willing to participate in. It is extremely difficult to collude on a small scale. An interesting point worth mentioning is that the active verifier set consists of the top verifiers with the most pledge tokens, and this threshold may be lowered in the future. In addition, each of their verifiers has stored the whole data set, and each node that stores a small amount of data may be optimized in the future, so if enough nodes are honest, the data can be reconstructed. The complete origin and more details can be found in the last post. This paper makes a beneficial comparison of the core components of the main layers, and there is a new discussion about the trade-off of each design in Ethereum. It is pointed out that a complex stack itself is not only pure, but also a set of smart contracts, but it depends on Ethereum, and otherwise, the Ethereum team thinks that tokens are necessary for security, and will eventually be needed because it is impossible to reduce the availability of data on the chain. They think that in order to ensure that nodes are honest, data can be. Using and punishing malicious nodes, the network must be able to be verified through the incentive structure including native tokens. Here, the criticism is made. Unless the source chain is bifurcated, the revalidator that keeps the data will not be verified. This is extremely impossible, because this is Ethereum. From the brand point of view, Ethereum is a product that is extremely consistent with Ethereum. According to the construction, the team is building it as the only data availability layer. He explained that by definition, the data availability layer is A modular product, but the others are actually the blockchain itself. Packaging the layers into the blockchain will indeed bring obvious benefits to the local operation on it, mainly in the form of security guarantee. However, it is mentioned that the goal of his team's construction is to create a product that provides data availability services for the Ethereum ecosystem from the first principle, a real one adjacent to the Ethereum ecosystem. He pointed out that there is no need for a separate consensus here because the Ethereum-based network has relied on sorting and consensus, which was explained in the recent program. Using validity proof and construction can achieve a high degree of flexibility in the ecosystem. 比特币今日价格行情网_okx交易所app_永续合约_比特币怎么买卖交易_虚拟币交易所平台
注册有任何问题请添加 微信：MVIP619 拉你进入群
打开微信扫一扫
添加客服
进入交流群
1.本站遵循行业规范，任何转载的稿件都会明确标注作者和来源；2.本站的原创文章，请转载时务必注明文章作者和来源，不尊重原创的行为我们将追究责任；3.作者投稿可能会经我们编辑修改或补充。
