网络地址转换：互联网中心化的根本性错误

网络地址转换（NAT）技术被指为互联网中心化的根本推手。1 RFC 1631在1994年正式引入了NAT，旨在解决IP地址枯竭和路由扩展两个核心问题。1 然而，这项原本作为短期应急方案的技术设计，却无意中限制了个人运行服务器的能力，训练用户接受客户端-服务器模式，最终推动了互联网向中心化和封闭生态系统的演进。1

为缓解NAT带来的限制，业界相继推出了端口转发、UPnP、STUN、TURN和ICE等解决方案。1 然而，运营商级NAT（CGNAT）的出现使用户完全无法控制端口转发，进一步强化了这一局面。1 原本IPv6被设想作为长期替代方案来彻底解决这一问题，但其采纳进程停滞不前。1 更为关键的是，NAT被错误地规范化为一种安全功能——通过"隐藏设备"的概念——这一误解阻碍了人们对替代方案的采纳，使得过时的技术架构得以持续存在。1

Network Address Translation (NAT) was formally proposed in RFC 1631 in 1994 as a short-term measure to address IP address exhaustion and routing scalability challenges 1. However, the technology's design has fundamentally shaped the structure of the modern internet in ways that extend far beyond its original technical scope 1.

By making it difficult for individuals to operate servers from their own networks, NAT conditioned users to accept a client-server model and outsource services to centralized providers 1. This architectural constraint has inadvertently accelerated internet centralization and the proliferation of walled gardens, as users became dependent on third-party platforms rather than hosting their own infrastructure 1. While IPv6 was envisioned as the long-term solution to address exhaustion, its adoption has stalled 1. Meanwhile, NAT has been improperly normalized as a security feature—the notion that "devices are hidden" behind NAT—which has cemented its continued use and discouraged the exploration of alternative approaches 1.

Technical workarounds including port forwarding, UPnP, STUN, TURN, and ICE have emerged to mitigate NAT's constraints 1. However, Carrier-Grade NAT (CGNAT), deployed by internet service providers, eliminates user control over port forwarding entirely, further entrenching the barriers to decentralized services 1. The persistence of NAT thus represents not merely a technical legacy, but a structural impediment to a more distributed internet architecture 1.