How Long Must a Secret Stay Secret?

A secret has a lifetime. A message arranging tomorrow's meeting might lose its value within days, while a research file could remain commercially sensitive for years. This difference matters when organisations prepare for quantum computers. The central question is not simply when a powerful new machine will arrive. It is whether information encrypted today must still be confidential when that machine becomes capable of attacking the protection around it.

秘密有自己的寿命。一条安排明天会议的消息，可能几天内就失去价值，而一份研究文件可能多年后仍涉及商业机密。当机构为量子计算机做准备时，这种差别十分重要。核心问题不只是强大的新机器何时出现，而是当这种机器有能力攻击信息的保护机制时，今天加密的信息是否仍然必须保密。

On 27 October 2026, the US National Institute of Standards and Technology, or NIST, announced a road map for moving to post-quantum cryptography. Its report explains the transition away from vulnerable algorithms towards replacement standards. NIST had released its first three finalised post-quantum standards in 2024. The new announcement concerns the difficult work of migration: deciding what must change, understanding different uses and organising the transition across systems and suppliers.

2026年10月27日，美国国家标准与技术研究院，简称NIST，公布了一份向后量子密码迁移的路线图。其报告说明了如何从存在脆弱性的算法过渡到替代标准。NIST已于2024年发布首批三项正式的后量子密码标准。这次公告涉及的是艰难的迁移工作：决定哪些内容必须改变，理解不同的使用方式，并协调跨系统、跨供应商的过渡。

Post-quantum cryptography uses mathematical methods intended to resist attacks from both conventional and quantum computers. It can run on ordinary computers; an organisation does not need to buy a quantum machine to use it. Nor does the announcement mean that today's quantum computers can already unlock everyone's messages. In a July 2026 explanation, NIST said current machines were too small and unstable to threaten cryptography. The arrival date of a sufficiently capable machine remains uncertain.

后量子密码采用旨在抵御传统计算机和量子计算机攻击的数学方法。它可以在普通计算机上运行；机构无需购买量子计算机就能使用它。这份公告也不意味着今天的量子计算机已经能够解开所有人的消息。在2026年7月的一篇说明中，NIST表示，目前的机器规模太小，而且不够稳定，尚不足以威胁密码技术。能力足够强大的机器何时出现，仍不确定。

The reason to act earlier is a threat called “harvest now, decrypt later”. An attacker could intercept encrypted information, retain a copy and wait for future capabilities to make it readable. Capturing a message and understanding it need not happen together. This separates the date of collection from the date of damage. A later security upgrade cannot reach back and remove a copy already held by an attacker. That is why the useful lifetime of a secret matters alongside the strength of its protection today.

需要提前行动的原因，是一种被称为“现在收集、以后解密”的威胁。攻击者可能截获加密信息，保留副本，等待未来的技术能力使其变得可读。获取一条消息和理解它的内容，不必同时发生。这使收集信息的日期与造成损害的日期分离。后来的安全升级无法回到过去，删除攻击者已经持有的副本。因此，秘密保持价值的时间，与它今天受到多强的保护，同样重要。

Consider a hypothetical engineering company choosing between two projects. One system sends routine delivery updates whose contents quickly become unimportant. Another exchanges designs that the company expects to keep secret for fifteen years. Assume both use cryptography vulnerable to a sufficiently powerful quantum computer. Looking only at today's traffic volume might put the delivery system first. Looking at confidentiality could reverse the order. The designs might still be valuable when a future attacker gains the ability to read captured messages. These are invented circumstances, not a reported company case.

考虑一家假设的工程公司，它正在两个项目之间确定优先顺序。一个系统发送日常配送更新，其中的内容很快就会变得无关紧要。另一个系统交换设计方案，公司预计需要将其保密十五年。假设两者使用的密码技术都无法抵御能力足够强大的量子计算机。只看今天的通信量，可能会把配送系统放在首位。考虑保密需求，则可能颠倒这个顺序。当未来的攻击者获得读取已截获消息的能力时，这些设计方案可能仍然有价值。这些情境是虚构的，并非报道中的企业案例。

There is another clock: the time needed to complete the upgrade. Suppose the design service depends on a supplier's software, a customer's older equipment and an internal application maintained by a different team. Changing one component would not demonstrate that the entire exchange was protected. The teams would need to establish which parts support the replacement methods and whether they remain compatible. This hypothetical dependency explains why an apparently small service could require early planning, even if another, larger service is easier to update.

还有另一只时钟：完成升级所需的时间。假设设计服务依赖一家供应商的软件、某位客户的旧设备，以及由另一支团队维护的内部应用。改变一个组件，并不能证明整个信息交换过程都得到了保护。各团队需要确定哪些部分支持替代方法，以及它们能否继续兼容。这种假设的依赖关系解释了，为什么一个看似很小的服务可能需要提前规划，即使另一个规模更大的服务反而更容易更新。

NIST's announcement specifically stresses the cryptographic supply chain. Its broader educational guidance also recommends identifying applications that use encryption and talking to technology departments and vendors. A useful inventory therefore supports questions, rather than merely producing a list: What information passes through this service? How long must it remain secret? Who controls the necessary changes? In planning terms, a system with a long preparation period deserves attention before its final installation date approaches. Priority involves both the consequences of exposure and the obstacles to completing the work.

NIST的公告特别强调了密码技术供应链。其更广泛的科普指南也建议识别使用加密的应用，并与技术部门及供应商沟通。因此，一份有用的清单应当帮助回答问题，而不仅仅是列出项目：哪些信息经过这项服务？它们必须保密多久？必要的变更由谁控制？从规划角度看，准备周期长的系统应当在最终安装日期临近之前就受到关注。优先顺序既涉及信息暴露的后果，也涉及完成工作的障碍。

This makes migration a scheduling problem under uncertainty. Waiting for a precise prediction about quantum hardware leaves the organisation's own preparation time untouched. Starting an inventory, identifying dependencies and testing a proposed change can reduce uncertainty that the organisation actually controls. Success would mean more than announcing that a new algorithm is available. It would mean being able to explain which information has gained protection, which connections still depend on older methods, and why the remaining work has been placed in that order.

这使迁移成为一个在不确定条件下安排进度的问题。等待有关量子硬件的精确预测，并不会缩短机构自身所需的准备时间。开始盘点、识别依赖关系，以及测试拟议的变更，可以减少机构真正能够控制的不确定性。成功意味着的不只是宣布某种新算法已经可用，还意味着能够说明哪些信息获得了保护，哪些连接仍依赖旧方法，以及为什么把剩余工作安排成这样的顺序。