为什么Solana这次宕机了这么久?

Odaily星球日报Опубликовано 2023-02-28Обновлено 2023-02-28

Введение

官方将尽快公布调查结果。

上周末,Solana 又宕机了……Solana 基金会表示,团队正积极调查宕机原因,很快就会公布。Odaily星球日报也会持续跟踪报道。

以下是推特用户@DBCrypt 0 对 Solana 在上周末宕机近 20 个小时原因的个人解读,我们将其编译整理如下。

宕机起因

Solana 的正常运行时间状态显示, 2 月 25 日,该网络遭遇了持续 18 小时 50 分钟的中断,这是 Solana 2023 年的首次网络中断。Solana 链的验证者两次试图重启网络,以恢复服务。

CoinDesk 的一份报告显示,Solana 上几乎所有的链上活动都在 25 号午夜后不久停摆。Solana 的 Discord 服务器上的用户报告说,区块链意外地开始“分叉”,创建了多个相互冲突的链历史版本。

Solana 有着丰富的网络中断“黑历史”, 2022 年就经历了 11 次大型网络中断和 3 次小型网络中断,宕机时间从 1 小时 15 分钟到 17 小时 7 分钟。而最近这次宕机是一年多来持续时间最长的。

@DBCrypt 0 表示,宕机是由于“一个巨大的设计缺陷”导致系统陷入困境,大量验证者消息和链上投票正在阻塞网络。他还解释说,链上共识使 Solana 陷入困境,当网络出现故障时,验证者通信将转移到 Discord,这可能会导致长时间的停机。

Solana 的设计缺陷

Solana 采用链上共识模型,这意味着网络交易由验证者和交易本身之间的共识通信组成——例如代币传输和铸造。(@DBCrypt 0 表示,这夸大了交易量。)

下面的图表显示了网络交易的快照。粉红色部分表示实际交易,而浅蓝色部分表示验证者验证通信。@DBCrypt 0 评论说,验证者通信消息占交易的 90% -95% ,这很“疯狂”。

“所以当 Solana 提到他们达到了 4000 TPS 时,只有不到 10% 是网络上的实际交易。”

来源:来自 @DBCrypt0 推特

由于大部分交易量是由验证者通信消息组成的,这“使系统陷入困境”。@DBCrypt 0 说,当网络中断时,验证者之间无法相互沟通。

在这种情况下,验证者会求助于 Discord,决定下面该怎么做。问题是,在任何提议的操作发生之前,必须得到三分之二的验证者的同意,而且有些验证者可能离线,还不知道出现宕机情况。

Solana 网络与 Hedera 网络的相似性

@DBCrypt 0 表示,Solana 通过操作链上共识模型“为全节点创建了大量的数据”。他补充说,由于历史记录中充斥着不必要的验证者消息,因此需要“数据中心”来运行完整的 Solana 节点。

@DBCrypt 0 最后指出,Hedera 网络也运行链上共识模型,并且与 Solana 一样存在固有的膨胀缺陷。

“很抱歉扫了所有的 Hedera 支持者的兴,但是这其中大多数交易都是不必要的,就像 Solana 一样。HBAR 代币的交易平均效率只有 3-5 TPS。”

2 月 25 日,SOL 遭受了 9% 的下跌,但在第二天的收盘价高于前一天的开盘价后反弹,这表明市场接受了 Solana 宕机的预期行为。

Solana 最新回应

Solana Labs 创始人兼首席执行官 Anatoly Yakovenko 回应称,@DBCryptoX 言论“纯属无知”。Yakovenko 指出,作为“单一庞大法定人数”的一部分,投票有助于同时提供“出色的安全性、高吞吐量和低费”。

然而,Yakovenko 并没有完全反驳 @DBCryptoX 这方面的说法,即“Solana 上 90-95% 的交易都包含这些验证者消息和链上投票,容易使系统陷入困境”。 此外,@DBCryptoX 还声称 Solana 网络中断持续 20 个小时,因为验证者需要相当长的时间才能使用链下方式(例如 Discord 等消息系统)沟通并达成共识(或达成解决方案)。

Solana 生态钱包 Phantom 软件工程师 Alex Kroeger 则表示,网络中断的原因可能并不单一,PoS 系统的验证者需要大量的网络通信来实现验证。

Похожее

After Marvell's 32% Surge, the Chinese Chip Family Behind It Emerges

The stock price of Marvell Technology surged 32.5% on June 2nd, driven by NVIDIA CEO Jensen Huang highlighting its custom ASICs and optical interconnects as core to AI data center architecture. This event brought attention to the Chinese semiconductor family behind Marvell: the Dai siblings. The story centers on three siblings, all UC Berkeley graduates, whose three-decade entrepreneurial journey aligns with major semiconductor industry shifts. In 1995, youngest sister Dai Wei Li co-founded Marvell with her husband Sehat Sutardja and his brother, focusing on storage controllers. Eldest brother Dai Wei Min founded EDA company Ultima, later sold to Cadence, and later founded VeriSilicon (芯原) in China, becoming a leading semiconductor IP provider. Second brother Dai Wei Jin co-founded EDA firm Silicon Perspective (sold to Cadence) and GPU IP company Vivante, later acquired by VeriSilicon. The combined "Dai-Sutardja" family network extends beyond Marvell. Their ventures and investments form a comprehensive ecosystem for the post-Moore's Law, chiplet era. Key holdings include: Dream Big Semiconductor (AI SuperNICs, acquired by Arm), Alphawave (high-speed SerDes IP, acquired by Qualcomm), and Silicon Box (a chiplet advanced packaging foundry). VeriSilicon itself thrives on the AI ASIC and IP boom in China. Collectively, the family's AI infrastructure-related portfolio is estimated at over $22 billion. Their strategy represents a distinct path: building critical components for open standards and key manufacturing capacity in the chiplet era, rather than pursuing standalone AI chip dominance. While this path may not create the next NVIDIA, it has enabled repeated successful exits and sustained influence within the global semiconductor industry.

marsbit39 мин. назад

After Marvell's 32% Surge, the Chinese Chip Family Behind It Emerges

marsbit39 мин. назад

Microsoft is Afraid of Being Marginalized by AI Giants

Microsoft, once the defining force of the PC era, now faces a familiar challenge in the AI age: the risk of being relegated to a profitable but invisible infrastructure provider. This anxiety was laid bare at Build 2026, where CEO Satya Nadella unveiled a major strategic pivot. The catalyst was a quiet April agreement that dissolved Microsoft's exclusive licensing and cloud-hosting deal with OpenAI, its once-vital partner. This erased Microsoft's key AI moat. With OpenAI and Anthropic defining AI applications and gaining enterprise traction—even within Microsoft's own ranks—Nadella had to answer: without exclusivity, what is Microsoft's role? The answer was a suite of seven in-house AI models, a developer-focused AI workstation (Surface RTX Spark Dev Box), and, most crucially, the Agent 365 platform for enterprise AI governance. The models, notably targeting Anthropic's strengths in coding and enterprise, signal a defensive move. However, the broader strategy is to make the models themselves less decisive. Financially, Microsoft's AI revenue is strong, driven largely by Azure running others' models. Yet its user-facing products like Copilot show weak penetration and engagement. Microsoft earns infrastructure money but lacks direct user mindshare. Nadella's core fear is being "hollowed out." As OpenAI and Anthropic prepare for IPOs and gain financial independence, they may build their own infrastructure, threatening Azure's lucrative AI revenue stream. Microsoft's window is to entrench itself deeper: not as the model creator, but as the indispensable platform for securely deploying, managing, and governing all AI models within the enterprise through Agent 365. Build 2026 revealed Microsoft's bet: in the AI era, the ultimate power lies not in any single model, but in the enterprise "operating system" that controls them. Nadella is determined to ensure Microsoft is the driver of this new era, not just a passenger.

marsbit53 мин. назад

Microsoft is Afraid of Being Marginalized by AI Giants

marsbit53 мин. назад

CPU, Quietly Returning to the Center of the AI Computing Power Stage

Over the past three years, AI computing power narratives have been dominated by GPUs. However, starting in 2026, this story began to shift. While training large models remains GPU-intensive, the rapid growth of inference and AI agent workloads, which require high levels of task orchestration, concurrency, and data flow management, has highlighted a renewed critical role for CPUs. These are tasks GPUs are not designed to handle. Intel's recent launch of the Xeon 6+ processor, built on its Intel 18A process and featuring up to 288 efficiency cores (E-cores), exemplifies this strategic pivot. It is positioned not as a mere companion to GPUs but as the essential "control plane" for AI infrastructure, optimized for high-density, energy-efficient, and high-throughput workloads characteristic of AI agents and inference. This "CPU resurgence" is not about CPUs outperforming GPUs in raw computation. It reflects a systemic bottleneck: as AI scales from training single models to deploying countless intelligent agents, the demand for coordination and data handling surges. Major cloud providers are also developing their own high-density ARM-based server CPUs for similar workloads. However, Intel's success with this strategy faces significant challenges. Competition includes NVIDIA's integrated CPU-GPU solutions, the expanding adoption of cloud vendors' in-house ARM CPUs, and the crucial market test of Intel's 18A manufacturing process against rivals like TSMC's N2. In conclusion, CPUs are indeed reclaiming a central, though redefined, role in AI compute—managing the complex orchestration that enables massive-scale AI deployment. While the trend is clear, which company will ultimately lead this CPU resurgence remains an open question to be decided in the data centers of 2027 and beyond.

marsbit1 ч. назад

CPU, Quietly Returning to the Center of the AI Computing Power Stage

marsbit1 ч. назад

Торговля

Спот
Фьючерсы

Популярные статьи

Как купить HBAR

Добро пожаловать на HTX.com! Мы сделали приобретение Hedera Hashgraph (HBAR) простым и удобным. Следуйте нашему пошаговому руководству и отправляйтесь в свое крипто-путешествие.Шаг 1: Создайте аккаунт на HTXИспользуйте свой адрес электронной почты или номер телефона, чтобы зарегистрироваться и бесплатно создать аккаунт на HTX. Пройдите удобную регистрацию и откройте для себя весь функционал.Создать аккаунтШаг 2: Перейдите в Купить криптовалюту и выберите свой способ оплатыКредитная/Дебетовая Карта: Используйте свою карту Visa или Mastercard для мгновенной покупки Hedera Hashgraph (HBAR).Баланс: Используйте средства с баланса вашего аккаунта HTX для простой торговли.Третьи Лица: Мы добавили популярные способы оплаты, такие как Google Pay и Apple Pay, для повышения удобства.P2P: Торгуйте напрямую с другими пользователями на HTX.Внебиржевая Торговля (OTC): Мы предлагаем индивидуальные услуги и конкурентоспособные обменные курсы для трейдеров.Шаг 3: Хранение Hedera Hashgraph (HBAR)После приобретения вами Hedera Hashgraph (HBAR) храните их в своем аккаунте на HTX. В качестве альтернативы вы можете отправить их куда-либо с помощью перевода в блокчейне или использовать для торговли с другими криптовалютами.Шаг 4: Торговля Hedera Hashgraph (HBAR)С легкостью торгуйте Hedera Hashgraph (HBAR) на спотовом рынке HTX. Просто зайдите в свой аккаунт, выберите торговую пару, совершайте сделки и следите за ними в режиме реального времени. Мы предлагаем удобный интерфейс как для начинающих, так и для опытных трейдеров.

693 просмотров всегоОпубликовано 2024.03.29Обновлено 2026.06.02

Как купить HBAR

Обсуждения

Добро пожаловать в Сообщество HTX. Здесь вы сможете быть в курсе последних новостей о развитии платформы и получить доступ к профессиональной аналитической информации о рынке. Мнения пользователей о цене на HBAR (HBAR) представлены ниже.

活动图片