Design object: the Internet of things the backend server
Design goal: 100000 concurrency
Programming languages: C/C + +
The working process:
100000 terminal to the server to send a status information every 3 seconds (ID, time, temperature, humidity, location, etc.), after you receive the back-end server in the server,
Users can through the client queries the historical data of a certain device, the user concurrency value below 50000,
This server should be how to design? A single computer can cope with?
My initial assumption is that
Terminal equipment - & gt; Server: MQTT + multi-threaded read + memory database buffer + + + mysql database connection (multithreaded) save
The client - & gt; Server: libUV + multithreading + + mysql database buffer memory read/write separation
.

You have a good plan to share, or what is the framework directly to share

CodePudding user response:

100000 terminals does not represent a concurrency value is 100000, concurrency value may be only a few hundred or even less,

CodePudding user response:

High concurrency Erlang, on the go,,,, why in c + + pit,

CodePudding user response:

Because of his server should be written in c + +

CodePudding user response:

reference 1/f, play big shoot early nuclear response:

100000 terminals does not represent a concurrency value is 100000, concurrency value may be only a few hundred or even less,

Libevent + the rabbitmq architecture can make high concurrency server?
Want to design a backend server architecture,
In front of the high concurrency this piece I'm going to use libevent to monitor and receive the request from the client, then the request message into the rabbitmq queues,
Then the worker thread get request from the queue to have a rest and treatment, would you please tell me what the problem is not so?

CodePudding user response:

refer to the second floor taodm response:

high concurrency Erlang, on the go,,,, why in the pit, c + +

Libevent + the rabbitmq architecture can make high concurrency server?
Want to design a backend server architecture,
In front of the high concurrency this piece I'm going to use libevent to monitor and receive the request from the client, then the request message into the rabbitmq queues,
Then the worker thread get request from the queue to have a rest and treatment, would you please tell me what the problem is not so?