Pub/Sub pattern 发布/订阅模式
发布订阅模式:发布方不用管理发布给哪个订阅方(看到这句话感觉不太好),有2种场景
场景2比较常见,多客户端向一个服务端订阅,场景1类似与REP/REQ的多服务器模式,避免单个PUB撑不住,(2个PUB的数据都能收到)
pub-server.py
port = '5556'
pub_server_name = 'pub-server01'
context = zmq.Context()
socket = context.socket(zmq.PUB)
socket.bind('tcp://*:%s'%port)
while True:
topic = random.randrange(9999,10005)
messagedata = random.randrange(1,215)-80
print 'topic:%s messagedata:%s'%(topic,messagedata)
socket.send('%d %d %s'%(topic,messagedata,pub_server_name))
time.sleep(1)
随机发送主题为9999-10005的数据
sub-client.py
#coding=utf-8
import sys
import zmq
port = "5556"
if len(sys.argv) > 1:
port = sys.argv[1]
int(port)
if len(sys.argv) > 2:
port1 = sys.argv[2]
int(port1)
# Socket to talk to server
context = zmq.Context()
socket = context.socket(zmq.SUB)
print "Collecting updates from weather server..."
socket.connect ("tcp://localhost:%s" % port)
if len(sys.argv) > 2:
socket.connect ("tcp://localhost:%s" % port1)
# 本地过滤
topicfilter = "10001"
socket.setsockopt(zmq.SUBSCRIBE, topicfilter)
# Process 5 updates
total_value = 0
for update_nbr in range (5):
string = socket.recv()
topic, messagedata,pub_server_name = string.split()
total_value += int(messagedata)
print topic, messagedata,pub_server_name
print "Average messagedata value for topic '%s' was %dF" % (topicfilter, total_value / update_nbr)
客户端只接收主题为10001的数据
PUB输出:
topic:10003 messagedata:7
topic:10000 messagedata:-27
topic:10000 messagedata:37
topic:10001 messagedata:-11
topic:10001 messagedata:-14
topic:9999 messagedata:38
client输出:
10001 -11pub-server01
10001 -14 pub-server01
另外:PUB/SUB是异步的,谁先开后开都行,就是说如果PUB已经开起来发送数据,然后SUB端起来,那么之前的那些数据SUB就收不到了,PUB和SUB逻辑上可以说是解耦的,二者不互相依赖。
最后是要点:
1.PUB未与SUB有逻辑上的连接,所以PUB可以很简单的drop掉数据(意思就是只管发)
2.如果你用的是TCP,然后SUB端要搞点什么比较慢,数据会Queue在PUB那头
3. 目前版本(最新版) filter(就是主题过滤)是在SUB端的,不是在PUB端,就是说SUB端比较苦逼,要全部收下一堆数据,再慢慢挑,这好像太那个了吧^_^