Tcp连接中可以分离一些数据吗? [英] Can some data be separated in a Tcp connection?

问题描述

我的问题很简单.可以在 tcp 连接中分离一些数据吗(如标题所说)?例如，如果我调用 MySocket.Send(Encoding.UTF8.GetBytes("this is a test!");，我可以从连接的另一端?即Tcp有多可靠?

my question is simple. Can some data be separated in a tcp connection (as title says)? For example if I call MySocket.Send(Encoding.UTF8.GetBytes("this is a test!");, may I get 2 separate receive operations like "this is" and " a test" from the other side of connection? Namely how much reliable is the Tcp?

推荐答案

TCP 通常用作流；不保证无论如何发送"和接收"操作将一对一匹配.必须读取 15 次，每次获取 1 个字节，这在技术上是有效的(但不太可能).因此，通常有两种操作模式:

TCP is usually used as a stream; no guarantee whatsoever is made that the "send" and "receive" operations will match up one-to-one. It would be technically valid (but unlikley) to have to read 15 times, getting 1 byte each time. As such, there are usually two modes of operation:

• 读取直到流结束(发送者在发送消息后关闭其出站套接字)
• 添加某种帧协议
  • 对于基于文本的协议，这通常是一个标记值，例如 0 或行尾(是 CR、LF、CRLF 还是 LFCR 取决于您！)
  • 对于二进制协议，这通常是一些定义明确的格式的长度前缀，告诉您下一条消息要考虑多少字节
  • read until the end of the stream (the sender closes their outbound socket after sending the message)
  • add some kind of framing protocol
    • for text-based protocols this is often a sentinel value, such as a 0 or an end-of-line (whether that is CR, LF, CRLF or LFCR is up to you!)
    • for binary protocols, this is often a length-prefix in some well-defined format, telling you how many bytes to consider for the next message

    (但这条线很模糊；例如，redis 协议大致基于文本，但使用了 ${n}\r\n{data}\ 形式的纯文本长度前缀r\n，其中 {n} 是表示为 ASCII 字符串的字节数，{data} 是原始的 n字节)

    (but the line is blurry; the redis protocol, for example, is roughly text-based, but uses a plain-text length-prefix of the form ${n}\r\n{data}\r\n, where {n} is the number of bytes expressed as an ASCII string, and {data} is the raw n bytes)

    这篇关于Tcp连接中可以分离一些数据吗?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！