11. application layer data transmission format / port number -bite

Design application layer protocol , It mainly includes two jobs : Specify the information to be transmitted and the organization format of the data ( Refer to the existing template :xml,json,protobuffer)

It is precisely because of the application layer protocol here , You can specify as you like , This leads to very serious polarization !!
The agreements designed by the boss are very good .
The agreement of vegetable chicken design is very bad
Because of that , The big guys have invented some better protocol templates , Let's go straight up :
Some popular templates of this kind of agreement ( Organization format of data )

Good readability , But the operation efficiency is not high :
1.xml
2. json
3. Poor readability , High operating efficiency :
4. protobuffer

1.xml:
Format consists of tag name :< Tag name > Content </ Tag name >
< Tag name >: The start tag
</ Tag name >: End tag

The tag name is key
The content is value

Improved readability : But too much supporting information is introduced , It takes up more network bandwidth

therefore , xml Now it is rarely used as the design template of application layer protocol , Now use xml Mainly as some configuration files .

2.json
adopt {} Key value structure . One {(} There are many key value pairs in . Use commas to separate key value pairs . Use colons between keys and values . There are many types of values that require that the key must be a string ( Numbers , character string , Boolean , Array , the other one json…)
{
key : value ,
key : value ,
…
}

json requirement key It must be a string , therefore key The quotation marks here can be omitted unless key It contains some special symbols ( Like spaces , perhaps - …) Must be quoted

json Represents a string , Single quotation marks or double quotation marks are OK ( Be similar to SQL)

The last key value pair , There can be a comma after , There can be no .( There are no standard requirements , But in general json Parser , Don't care about this detail )

Compared with xml, json It also ensures readability without xml So complicated ( Use less bandwidth )

however , When representing a more complex data , For example, there are many arrays here key Will repeat N Time , It takes up more extra bandwidth

3.protobuffer:
Is a binary format of data :

stay protobuffer Data in , No longer include the above key The name of . But through order and some special symbols , To distinguish the meaning of each field

At the same time, pass another IDL File to describe the data format ( What does each part mean ), IDL Just play an auxiliary development effect , It won't really transmit . Only binary pure data is transmitted

This is a simplified representation version :
1\3\3 Fried rice with egg \210\21\210\3 Stir-Fried Noodles with Vegetables \28\2\1\28 Rearrange the content here through binary data , There may even be some data compression .

In this way, the transmission efficiency will be higher , But it will also make this data difficult to observe with the naked eye , Debugging is inconvenient .

summary :
json The scope of application is wider than protobuffer More open
Hair efficiency > Operational efficiency

If something goes wrong with the online environment
If it is to use json, Requests and responses that go wrong , Be clear at a glance , If it is to use protobuffer, Binary data , Can't see with the naked eye , You have to develop a special program to analyze this data , Analyze the problem here

The most famous application layer protocol :HTTP

Port number

0-65535 Integer between

Well known port number : hold 0-1024 These port numbers , It has been divided into some specific functions , Many network services are very commonly used , For a wide range of services, special port numbers are assigned to these services for better management , It's not mandatory , It's just advice :

80 http The server
443 https The server
22 ssh
23 ftp
…