pyndk

python 高性能,高并发，轻量级，多进程，多线程，异步非阻塞网络开发框架。适用linux 下 python 2.6-2.7版本（不能用于windows，因为用了select.epoll ) 4w/qps

单进程，及多进程DEMO

    from pyndk.CCommMgr import *
    from pyndk.CRawPackageFilter import CRawPackageFilter
    from pyndk.CProcessor import CProcessor

    class ProCenter(CProcessor):
        instance = None
    
        @staticmethod
        def getInstance():
            if ProCenter.instance == None:
                ProCenter.instance = ProCenter()
            return ProCenter.instance
        def __init__(self):
            pass
        def init(self):  
            pass
         #收到完整包会调用此函数     
        def onRead(self, session, package): 
            response = "HTTP/1.1 200 OK\r\nContent-Length: 5\r\n\r\nhello"
            CCommMgr.getInstance().write(session,response,True);
        #客户端连接时会调些函数    
        def onConn(self, session, flag): 
            pass
            #print 'onConn'
        #客户端连接断掉会调用此函数    
        def onClose(self, session): 
            #print 'onClose'
            pass
         #网络IO出错会调用此函数    
        def onError(self, msg):
            print ('onError')
          #记时器回调  
        def onTimer(self, timerId, data): #
            print 'onTimer'
            CCommMgr.getInstance().setTimer(1, 1000, self, data)
            
    if __name__ == '__main__':
        #创建服务，监听9003端口，包过滤器为 CRawPackageFilter（用来确定包的完整性）
        srvId1 = CCommMgr.getInstance().createSrv(SRV_TCP, '0.0.0.0', 9003, CRawPackageFilter())
        #初始化处理器
        ProCenter.getInstance().init()
        #指定服务处理器
        CCommMgr.getInstance().setProcessor(srvId1, ProCenter.getInstance())
        #设置记时器处理器
        CCommMgr.getInstance().setTimer(1, 1000, ProCenter.getInstance(), None)
        #以单进程epoll事件模型启动服务
        #CCommMgr.getInstance().start() single process the better performance and worst concurrency
        #以多进程模epoll 事件模型启动服务
        CCommMgr.getInstance().fork(4) #multi process the better performance and better concurrency

ab测试性能 100 并发

    Concurrency Level:      100
    Time taken for tests:   2.429 seconds
    Complete requests:      100000
    Failed requests:        0
    Total transferred:      4300000 bytes
    HTML transferred:       500000 bytes
    Requests per second:    41168.41 [#/sec] (mean)
    Time per request:       2.429 [ms] (mean)
    Time per request:       0.024 [ms] (mean, across all concurrent requests)
    Transfer rate:          1728.75 [Kbytes/sec] received