官方给出的定义如下：

/**
 * \ingroup tcp
 * \brief TCP socket creation and multiplexing/demultiplexing
 * 
 * A single instance of this class is held by one instance of class Node.
 *
 * The creation of TcpSocket are handled in the method CreateSocket, which is
 * called by TcpSocketFactory. Upon creation, this class is responsible to
 * the socket initialization and handle multiplexing/demultiplexing of data
 * between node's TCP sockets. Demultiplexing is done by receiving
 * packets from IP, and forwards them up to the right socket. Multiplexing
 * is done through the SendPacket function, which sends the packet down the stack.
 *
 * Moreover, this class allocates "endpoint" objects (ns3::Ipv4EndPoint) for TCP,
 * and SHOULD checksum packets its receives from the socket layer going down
 * the stack, but currently checksumming is disabled.
 *
 * \see CreateSocket
 * \see NotifyNewAggregate
 * \see SendPacket
*/

简要来说即为：

该类的一个实例由一个Node类实例保存。

TcpSocket的创建在CreateSocket方法中处理，该方法由TcpSocketFactory调用。 在创建时，该类负责套接字初始化并处理节点的TCP套接字之间的数据复用/解复用。 通过从IP接收数据包完成解复用，并将它们转发到正确的套接字。 多路复用是通过SendPacket函数完成的，该函数将数据包发送到堆栈。

此外，此类为TCP分配“端点”对象（ns3 :: Ipv4EndPoint），并且应该从堆栈中的套接字层接收校验和数据包，但当前校验和已被禁用。

我们可以通过以下代码来设置拥塞控制算法：

Config::SetDefault ("ns3::TcpL4Protocol::SocketType",
                          TypeIdValue (TypeId::LookupByName (transport_prot)));

ns3::TcpL4Protocol::SocketType变量类型决定了TCP拥塞控制算法。即

We pick the TCP congesion-control algorithm by setting ns3::TcpL4Protocol::SocketType. Options are TcpRfc793 (no congestion control), TcpTahoe, TcpReno, TcpNewReno and TcpWestwood. TCP Cubic and SACK TCP are not supported natively (though they are available if the Network Simulation Cradle is installed).

参考文献