Stanford CS144 Introduction to Computer Networking 学习笔记（二）

1-5: The Principle of Packet Switching

Packet switching is the idea that we break our data up into discrete, self-contained chunks of data. Each chunk, called a packet, carries sufficient information that a network can deliver the packet to its destination.

设计理念

Packet switch(分组交换机): 不需要知道per-flow state，即packet属于哪个flow

而对应的在电话网络中应用的circuit switching，其switch是需要维护发送方-接收方connection state，并且为每个发送方-接收方维护一条连接链路

packet switching有两种方式：（1）self routing(不再使用)：packet的header中包含有它从source到destination的路线；（2）forwarding table: a switch can have a table of destination addresses and the next hop

Flow: A collection of datagrams belonging to the same end-to-end communication, e.g. a TCP connection.

好处

packet switch简单

多个flow可以充分利用link的能力(Statistical Multiplexing)

1-6: The Principle of Layering

layer只和上下层的layer交流，下层layer为上层提供了well-defined service，各layer可以随着时间改进而不影响其他层

举例：邮局系统；航空公司；编程

有时为了特殊目的，需要打破layering，比如在c语言中加入assembly code，使得上层和底层的processor不再相互独立。

1-7: The Principle of Encapsulation

Encapsulation allows you to layer recursively, 比如TCP segment可以在另一个TCP segment，应用VPN(virtual private network)

Encapsulation is how layering manifests in data representation

• Help separation of concerns

• Help enforce boundaries/layering

• Simplify layer implementations

1-8: Byte order and packet formats

Internet使用的协议都是big endian，即packet to be in a big endian format

network (byte) order is big endian，网络编程时需要将network order转为host order

文本数据在内存中的layout与字节顺序无关，在big endian和little endian的机器上显示结果相同