Cross platform high performance TCP network library and RPC library using C++ 11.
Windows : Linux :
- Cross platform (Linux | Windows)
- High performance and safety use.
- None depend
- Multi-threaded
- SSL support
- Support HTTP、HTTPS、WebSocket protocol
- IPv6 support
- RPC Library
Under localhost, use CentOS 6.5 virtual mahcine(host machine is Win10 i5)
-
PingPong
Benchamrk's server and client both only use one thread, and packet size is 4k
-
Broadcast
Server use two network threads and one logic thread, client use one network(also process logic) thread. every packet size is 46 bytes. every packet contain client's id. server broadcast packet to all client when recv one packet from any client. client send one packet when recv packet from server and packet's id equal self.
-
Ab HTTP(1 network thread)
Server Software: Server Hostname: 127.0.0.1 Server Port: 8080 Document Path: / Document Length: 18 bytes Concurrency Level: 100 Time taken for tests: 5.871 seconds Complete requests: 100000 Failed requests: 0 Write errors: 0 Non-2xx responses: 100000 Total transferred: 5200000 bytes HTML transferred: 1800000 bytes Requests per second: 17031.62 [#/sec] (mean) Time per request: 5.871 [ms] (mean) Time per request: 0.059 [ms] (mean, across all concurrent requests) Transfer rate: 864.89 [Kbytes/sec] received Connection Times (ms) min mean[+/-sd] median max Connect: 0 2 0.7 2 8 Processing: 1 3 0.7 3 9 Waiting: 0 3 0.8 3 8 Total: 2 6 0.8 6 11 Percentage of the requests served within a certain time (ms) 50% 6 66% 6 75% 6 80% 6 90% 7 95% 7 98% 7 99% 8 100% 11 (longest request)
This library use three layer ident one session(also is three way to use this library).
- First, use raw pointer named DataSocket, combine with EventLoop in used
- Second, use int64_t number ident one session, used in some callback of TCPService that wrapper DataSocket of first layer
- Thrid, use smart pointer named TCPSession::PTR combine with WrapServer, you can control session by TCPSession::PTR
I suggest you use the second or thrid way above, because don't worry memory manager
About TCP Library ,Please see examples;
Use this RPC library, you not need any proto file sush as Protobuf and Thrift, because i use C++ Template Generic Programming do this work.
you can click here see example.
The RPC support any C++ base type, sush as int、string、vector、map, and support Protobuf Message Type; Of course, RPC can use async callback mode, when you need process RPC reply msg return from server.
On server side:
static int add(int a, int b)
{
//CenterServerRPCMgr::getRpcFromer()为调用者会话对象
return a + b;
}
static void addNoneRet(int a, int b, dodo::RpcRequestInfo reqInfo)
{
// 添加dodo::RpcRequestInfo reqInfo形参(不影响调用者调用)
// 这里本身不返回数据(函数返回类型为void),但RPC本身是具有返回值语义的
// 适用于需要调用其他异步操作之后(通过reqInfo)才能返回数据给调用者的情况
// 譬如:
/*
auto caller = CenterServerRPCMgr::getRpcFromer();
redis->get("k", [caller, reqInfo](const std::string& value){
caller->reply(reqInfo, value);
});
*/
/*
//客户端:
centerServerConnectionRpc->call("test", 1, 2, [](const std::string& value){
});
*/
}
void initCenterServerExt()
{
CenterServerRPCMgr::def("test", [](int a, int b){
return a + b;
});
CenterServerRPCMgr::def("testNoneRet", [](int a, int b){
});
CenterServerRPCMgr::def("add", add);
CenterServerRPCMgr::def("addNoneRet", addNoneRet);
}
On client side:
gLogicCenterServerClient->call("test", 1, 2);
gLogicCenterServerClient->call("add", 1, 2);
gLogicCenterServerClient->call("add", 1, 2, [](int result) {
cout << result << endl;
});