东京——重创微软365（Microsoft 365）的国际服务中断事件，暴露了政府、企业和其他组织依赖单一供应商服务的风险，同时也凸显了大量数据集中在云系统中造成的脆弱性。
TOKYO -- The international outage that hit Microsoft 365 has exposed the risk of governments, businesses and other organizations becoming dependent on a single provider's services, while also highlighting the vulnerabilities created by concentrating large amounts of data in cloud systems.
“我们无法发送邮件确认交易细节，所以改用传真机，”日本一家信托银行的基金经理周五在服务仍中断时表示。
"We can't send emails to confirm transaction details, so we're using the fax machine instead," a fund manager at a Japanese trust bank said on Friday while the outage was still ongoing.
据追踪在线服务中断情况的网站Downdetector称，周五日本时间凌晨2点过后不久，问题报告激增。涉及微软主要服务的报告达到每分钟近19650条。
Reports of problems surged shortly after 2 a.m. Japan time on Friday, according to Downdetector, a website that tracks online service disruptions. Reports involving major Microsoft services reached nearly 19,650 per minute.
此次中断扰乱了日本的多种服务。快递公司佐川急便表示，其让旅客在线预订行李寄存的“佐川免提服务”无法使用。电子产品零售商Bic Camera表示，顾客无法使用其在线预订服务，该服务允许购物者在公司电商网站上预留商品，然后到店取货。
The outage disrupted various services in Japan. Delivery company Sagawa Express said its "Sagawa Hands-Free Services," which let travelers reserve baggage storage online, were unavailable. Electronics retailer Bic Camera said customers could not use its online reservation service, in which shoppers set aside products on the company's e-commerce site for pickup in stores.
使用微软Outlook电子邮件客户端软件的日本政府机构无法收发邮件，也无法使用Teams消息平台沟通。
Japanese government offices that use Microsoft's Outlook email client software were unable to send and receive emails and they could not communicate using the Teams messaging platform.
微软的IT服务是许多组织的核心业务基础设施，这意味着服务中断可能造成大范围混乱。
Microsoft's IT services function as core business infrastructure for many organizations, meaning outages can cause widespread disruption.
总部位于东京的电子邮件管理服务提供商Onebox的一项调查发现，微软365是使用最广泛的工作场所电子邮件平台，55%的受访者表示他们使用它。另有13%使用较旧的微软电子邮件软件，凸显了对该公司的依赖程度。涉及云服务的重大中断并不罕见。
Microsoft 365 is the most widely used workplace email platform, a survey by Tokyo-based email management service provider Onebox found, with 55% of respondents saying they use it. Another 13% use older Microsoft email software, underscoring the degree of dependence on the company. Major outages involving cloud services are not uncommon.
2024年，Salesforce用户无法访问该公司的服务。此次中断影响了市政灾害信息网站的访问，并干扰了金融机构的在线信用卡贷款申请。
In 2024, Salesforce users were unable to to the company's services. The disruption affected access to municipal disaster information websites and interfered with online credit-card loan applications at financial institutions.
同样，2025年10月亚马逊云服务（AWS）的一次故障影响了约2500家公司。AWS客户运营的网约车服务和游戏服务器一度无法访问。微软此次故障扰乱了全球企业和公共机构的运营。
Similarly, an outage at Amazon Web Services in October 2025 affected roughly 2,500 companies. Ride-hailing services and game servers operated by AWS customers became temporarily inaccessible. The Microsoft outage disrupted operations at companies and public institutions worldwide.
在新加坡，国内税务局和法院等政府机构警告称，电子邮件回复将会延迟，并敦促公众改用在线表格或电话咨询。
In Singapore, government agencies including the Inland Revenue Authority and the courts warned that email responses would be delayed and urged the public to use online forms or telephone inquiries instead.
泰国的曼谷银行在失去电子邮件功能后，改为通过电话处理客户沟通。国有银行泰京银行和暹罗商业银行也采取了类似措施。开泰银行则转向Line即时通讯应用来维持客户沟通。
Thailand's Bangkok Bank switched to handling customer communications by phone after losing email capabilities. State-owned Krungthai Bank and Siam Commercial Bank adopted similar measures. Kasikornbank turned to the Line messaging app to maintain customer communications.
影响范围广泛，反映出微软云端服务相互关联的结构。微软的服务状态页面显示，问题波及多个产品，包括Teams、电子邮件基础设施服务Exchange、人工智能助手Copilot和信息共享软件SharePoint。
The broad impact reflected the interconnected structure of Microsoft's cloud-based services. Microsoft's service-status page showed problems occurred across products including Teams, email infrastructure service Exchange, artificial intelligence assistant Copilot and information-sharing software SharePoint.
近年来，全球企业越来越多地采用生成式AI来提高生产力和开发新创意。Copilot可以访问存储在Teams和Outlook中的日程安排、会议纪要和其他数据，然后为用户总结和呈现这些信息。这种整合的便利性促使各组织将工作相关信息集中到微软生态系统中。但一旦发生故障，影响也会被放大。微软尚未披露此次故障原因的相关细节。
In recent years, companies worldwide increasingly have adopted generative AI to improve productivity and develop new ideas. Copilot can access schedules, meeting minutes and other data stored in Teams and Outlook, then summarize and present that information to users. The convenience of such integration has encouraged organizations to consolidate work-related information in the Microsoft ecosystem. But the impact is amplified when disruptions occur. Microsoft has not disclosed details regarding the cause of the outage.
“单台服务器能处理的流量是有限的，因此系统通常会通过控制网络路由，将流量分散到多台服务器上，”神户大学名誉教授、信息安全专家森正胜表示。关于此次最新宕机，“旨在防止流量集中于某一地点的保障措施可能失效了。”森正胜表示，这一教训不仅适用于微软。
"The amount of traffic a single server can handle is limited, so systems are normally designed to distribute traffic across multiple servers by controlling network routes," said Masakatsu Morii, professor emeritus at Kobe University and an expert on information security. Regarding the latest outage, "safeguards intended to prevent traffic from concentrating in one location may have failed." The lesson extends beyond Microsoft, Morii said.
“公司必须假设，无论规模大小，任何云服务都可能发生宕机，并保持备用的通信方式，”他说。
"Companies must assume that outages, large or small, are going to happen with any cloud service and maintain alternative means of communication," he said.