docs: update docs

src/en/guide/concepts/network/2-tcp-udp.md

@@ -81,11 +81,11 @@ If the client sends a `SYN` command, and drops the line before the server return
 
 :::
 
-##### Sequence numbers and acknowledgment signals
+##### **Sequence numbers and acknowledgment signals**
 
 TCP uses sequence numbers to sort and deduplicate packets, and through the ACK acknowledgment mechanism, it ensures that packets are successfully delivered, ensuring the integrity of the data.
 
-##### Timeout retransmission
+##### **Timeout retransmission**
 
 TCP can retransmit packets until an ACK acknowledgment is received when packets are lost or delayed, through the timeout retransmission mechanism.
 
@@ -137,9 +137,59 @@ The SACK method ([RFC 2018](https://www.ietf.org/rfc/rfc2883.txt)) adds a buffer
 D-SACK ([RFC 2883](https://datatracker.ietf.org/doc/html/rfc2018)) mainly solves the problem of ACK loss. D-SACK uses the first segment of SACK as a flag to mark the packets that have been ACKed. When the receiving end receives a duplicate packet, it will write the packet into the D-SACK flag, telling the sender that the packet has been received, and the ACK may have been lost.
 :::
 
-#### Checksum
+##### **Flow Control**
 
-TCP also calculates a checksum on the data to ensure that the data was not lost or corrupted during transmission.
+TCP uses a sliding window to control traffic, allowing the sender to control the speed of data transmission based on the receiver's receiving capacity.
+
+::: warning 
+The receiver will write the currently available window size into the TCP header when responding with `ACK`, and the sender adjusts the transmission rate based on the window size.
+:::
+
+##### **Congestion Control**
+
+::: tip Algorithms
+- Slow Start
+- Congestion Avoidance
+- Fast Retransmit
+- Fast Recovery
+:::
+
+![Congestion Control](/assets/image/article/network/tcp-congestion-control.png)
+
+::: info Related Terms
+- cwnd - Congestion Window
+- ssthresh - Slow Start Threshold
+:::
+
+::: important
+
+- If `cwnd < ssthresh`, use the Slow Start algorithm.
+
+- If `cwnd > ssthresh`, use the Congestion Avoidance algorithm.
+
+- If `cwnd = ssthresh`, use either the Slow Start or Congestion Avoidance algorithm.
+:::
+
+###### **Slow Start**
+
+In the Slow Start phase, the initial value of `cwnd` is `1`, and `cwnd` will double every time a propagation round passes, until it reaches `ssthresh`.
+
+###### **Congestion Avoidance**
+
+When `cwnd` reaches `ssthresh`, the Congestion Avoidance algorithm will be executed. At this time, the growth of `cwnd` changes from doubling to adding `1` every time a propagation round passes. If the sender detects network congestion (i.e., the sent message does not get a timely response), it will set `ssthresh` to half of the `cwnd` value when network congestion occurs, and the value of `cwnd` will be reset to `1`, and Slow Start will be re-executed.
+
+###### **Fast Retransmit and Fast Recovery**
+
+Fast Retransmit has been introduced earlier, that is, if three consecutive duplicate acknowledgments are received, the yet-to-be-received packet segment will be sent immediately. The Fast Recovery algorithm needs to be used in conjunction with the Fast Retransmit algorithm.
+
+::: info Fast Recovery Algorithm
+- When the sender receives three consecutive duplicate acknowledgments, it halves `ssthresh`.
+- Set `cwnd` to the size of `ssthresh`
+:::
+
+##### **Checksum**
+
+TCP will also calculate a checksum for the data to ensure that the data is not lost or erroneous during transmission.
 
 ### UDP
 

src/guide/concepts/network/2-tcp-udp.md

@@ -81,11 +81,11 @@ TCP 建立连接的过程就是同步序列号的过程，SYN (Synchronize Seque
 
 :::
 
-##### 序列号和确认应答信号
+##### **序列号和确认应答信号**
 
 TCP 通过序列号可以对数据包进行排序和去重，同时通过 ACK 应答机制确保数据包成功送达，保证了数据的完整性。
 
-#### 超时重传
+##### **超时重传**
 
 TCP 通过超时重传机制可以在数据包丢失或者延迟时，重新发送数据包直到收到 ACK 应答。
 
@@ -137,7 +137,7 @@ SACK 方法 ([RFC 2018](https://www.ietf.org/rfc/rfc2883.txt)) 在 tcp 的 optio
 D-SACK ([RFC 2883](https://datatracker.ietf.org/doc/html/rfc2018)) 主要解决了 ACK 丢失的问题， D-SACK 使用 SACK 第一个段作为标志位，用于标记已经 ACK 的数据包。 当接收端接收到重复的报文段时会将该报文段写入 D-SACK 的标志位，告诉发送方已接收到报文段， ACK 可能已丢失。
 :::
 
-#### 流量控制
+##### **流量控制**
 
 TCP 使用滑动窗口来控制流量，使得发送端可以根据接收端的接收能力控制发送数据的速度。
 
@@ -148,7 +148,7 @@ TCP 使用滑动窗口来控制流量，使得发送端可以根据接收端的
 接收端在应答 `ACK` 的时候会将当前可用的窗口大小写入 tcp 头部，发送端根据窗口大小调整发送速率。
 :::
 
-#### 拥塞控制
+##### **拥塞控制**
 
 ::: tip 算法
 - 慢开始
@@ -173,15 +173,15 @@ TCP 使用滑动窗口来控制流量，使得发送端可以根据接收端的
 - `cwnd = ssthresh`，使用慢开始与拥塞避免算法任意一个。
 :::
 
-##### 慢开始
+###### **慢开始**
 
 在慢开始阶段， `cwnd` 初始值为 `1`, 每经过一个传播轮次，`cwnd` 都会翻倍，直到达到 `ssthresh` 。
 
-##### 拥塞避免
+###### **拥塞避免**
 
 当 `cwnd` 达到 `ssthresh` ，则会执行拥塞避免算法，此时 `cwnd` 的增长从翻倍增长变为每经过一个传播轮次就加 `1` 。若发送方检测到网络拥塞（即发送的消息没得到及时的回应），就会将 `ssthresh` 设置为发生网络拥塞时的 `cwnd` 值的一半，同时 `cwnd` 的值将会重置为 `1` ，重新执行慢开始。 
 
-##### 快重传和快恢复
+###### **快重传和快恢复**
 
 快重传在前面已介绍过，即如果连续收到三个重复的确认就会立即发送尚未接收到的报文段。快恢复算法需要配合快重传算法使用。图中的 5 号为快恢复阶段。
 
@@ -190,7 +190,7 @@ TCP 使用滑动窗口来控制流量，使得发送端可以根据接收端的
 - 将 `cwnd` 设置为  `ssthresh` 的大小
 :::
 
-#### 校验和
+##### **校验和**
 
 TCP 还会对数据计算校验和，确保数据在传输过程中未丢失或出现错误。