自己实现的一个源代码来说明UDP穿越NAT的原理。 首先先介绍一些基本概念： NAT(Network Address Translators)，网络地址转换：网络地址转换是在IP地址日益缺乏的情况下产生的，它的主要目的就是为了能够地址重用。NAT分为两大类，基本的NAT和NAPT(Network Address/Port Translator)。最开始NAT是运行在路由器上的一个功能模块。 最先提出的是基本的NAT，它的产生基于如下事实：一个私有网络（域）中的节点中只有很少的节点需要与外网连接（呵呵，这是在上世纪90年代中期提出的）。那么这个子网中其实只有少数的节点需要全球唯一的IP地址，其他的节点的IP地址应该是可以重用的。因此，基本的NAT实现的功能很简单，在子网内使用一个保留的IP子网段，这些IP对外是不可见的。子网内只有少数一些IP地址可以对应到真正全球唯一的IP地址。如果这些节点需要访问外部网络，那么基本NAT就负责将这个节点的子网内IP转化为一个全球唯一的IP然后发送出去-He s a source to explain the principles through NAT UDP). First introduced some basic concepts: NAT (Network), its "Network Address translation: Network Address translation is growing in the IP Address of the circumstances, its main purpose is to Address reuse. NAT is divided into two categories, the basic NAT and NAPT Network (" Port/Translator). Most started on a router NAT is running a function module. The first is the basic NAT, which is based on the fact that, a private network (domains) nodes in only a few nodes need and connected to the network (ha ha, this is in the mid 1990s put). So the subnet of only a few of the nodes in the only global IP address to other nodes IP address should be reused. Therefore, the basic functions of the NAT is very simple, use a reserve in subnet mask, the IP is not visible outside the IP. Subnet only a few IP address can correspond to the real world of IP addresses. If the nodes need access to external network, then basic NAT is responsible for thi