LoadRunner tests the application system of the C/S architecture (turn: how to test the C/S performance)

It has been a long time for me to use LoadRunner to test the C/S architecture software and how to select the protocol. I have read this article, and I hope it will be useful to everyone.

First, what is a protocol?
The Protocol is nothing more than an agreement. The agreement on the format of data packet sending means that if everyone sends data in this way, the communication will succeed, then there is no way to communicate.
The next step is how LR works during recording. The recording of LR is different from that of WR. It does not care about your object recognition or window. lr has an agent process, to monitor the communication between the client and the server, and then use your own function for recording. Therefore, when recording LR, we care about communication, which is the data packet between the client and the server. Here, we know clearly why we cannot record data sometimes? Because the protocol is unknown, the packets intercepted by LR cannot be parsed, so the recording is empty.
Now let's take a look. How can we select a protocol? The principle is that the communication protocol of your data packets can be recognized by LR.
A popular saying in the past is that as long as the B/S structure is based on the HTTP protocol, if it is not B/S, it must be socket. In fact, this saying is superficial or one-sided, I think to really understand this problem, you must understand what protocol packages are used for the data flow of the system you are measuring. I personally think it is best to learn more from developers. (Speaking of this, I would like to suggest one thing: it is a good habit for testers to learn from developers, to learn more about the underlying things, or to program architecture and data flow, I have a better understanding of internal structure analysis, which is helpful for my own tests and for my own growth.) In addition, I personally think that, as a tester, you need to learn more about the network. It is best to learn some network analysis tools, such as Sniffer, which is helpful for testing.
After talking so much, it seems that the question is gone? Back to the question, how to select the protocol.
I would like to recommend some suggestions below. I have seen them in a professional test book and posted them for your reference only. I am still talking about the specific analysis of the specific issue. Choosing an agreement is not a matter of dogma, but a matter of research and exploration.
Protocol selection reference:
Application Type protocol selection
1. Web site http/html
2. FTP Server
3. Email Server IMAP, POP3, SMTP
4. C/S (first) Client Connected to the background database ms SQL Server, Oracle, Sybase, DB2, infrmix using ADO and oledb Methods
C/S (second type) Client connects to the background database ODBC using ODBC Method
C/S (third type) No background database socket
5. ERP system sap lelesoft
6. Distributed Component COM/dacom EJB
7. Wireless Application WAP palm

In short, only by fully understanding the application type and technical architecture of the tested system can you make the right choice.

Still in confusion, you can only continue research ....