分组交换：
一种数据传输模式，在该模式中，消息被分成多个独立发送的部分，通过每个数据包的最佳路由，并在目的地重新组合。分组交换在交换分组时使用存储和转发技术；在转发数据包时，每跳首先存储该数据包然后转发。
众所周知，数据被分成许多小包以减少总延迟。所以你怎么看?数据包数量越大，延迟越小?
不，实际上这不是真的。好的，让我们通过例子来理解。
例子：
Given,
Data size = 1000 Bytes
Bandwidth = 1 MBps
Header size = 100 Bytes
Number of Hops = 3
assume Propagation dalay = 0
说明：以下是不同的场景。
案例 1：如果数据包数 = 1
Size of the Data Packet,
= Data size + Header size
= 1000 + 100 = 1100 Bytes
Transmission delay
= L/BW = 1100/10^6 = 1.1 ms
Total time taken
= number of Hopes * Transmission delay
= 3 * 1.1 = 3.3 ms 案例 2：如果数据包数 = 5
Size of each Data Packet
= (1000/5) + 100 = 300 Bytes
Transmission Delay for each Packet
= 300/10^6 = 0.3 ms 
Time taken by 1st Packet,
= No of Hops * Transmission delay
= 3*0.3 = 0.9 ms
Time taken by the remaining 4 Packets
= 4*Transmission delay
= 4*0.3 = 1.2 ms
Total Time Taken,
= 0.9+1.2 = 2.1 ms 案例 3：如果数据包数 = 10
Size of each Data Packet
= (1000/10) + 100 = 200 Bytes
Transmission delay for each packet
= 200/10^6 = 0.2 ms
Time Taken by first Packet
= 3*0.2 = 0.6 ms
Time Taken by remaining Packets
= 9*0.2 = 1.8ms
Total Time Taken
= 0.6+1.8 = 2.4 ms 案例 4：如果数据包数 = 20
Size of each Data Packet
= (1000/20) + 100 = 150 Bytes
Transmission delay for each Packet
= 150/10^6 = 0.15 ms
Time Taken by 1st Packet
= 3*0.15 = 0.45 ms
Time Taken by remaining 19 Packets
= 19*0.15 = 2.85 ms
Total Time Taken
= 0.45 + 2.85 = 3.3 ms 正如我们所看到的，存在一个总时间减少的阈值，如果我们在该限制之后增加数据包数量，总时间开始增加。最初时间会减少一些限制，之后时间会增加。如果数据包的数量非常大，那么它所花费的时间可能比传输单个数据包所花费的时间更长。