一.实验代码node
import httplib2 import time import json class OdlUtil: url = '' def __init__(self, host, port): self.url = 'http://' + host + ':' + str(port) def install_flow(self, container_name='default',username="admin", password="admin"): http = httplib2.Http() http.add_credentials(username, password) headers = {'Accept': 'application/json'} flow_name = 'flow_' + str(int(time.time()*1000)) #s2流表 #h2工做时s2端口1流量空闲时下发的流表 h2_to_s2_1 ='{"flow": [{"id": "0","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.2/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "1"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "1","table_id": "0"}]}' #h3工做时s2端口1流量空闲时下发的流表 h3_to_s2_1 ='{"flow": [{"id": "1","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.3/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "1"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "1","table_id": "0"}]}' mh3_to_s2_1 ='{"flow": [{"id": "1","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.3/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "1"},"order": "0"}]}}]},'\ '"priority": "100","cookie": "1","table_id": "0"}]}' #h3工做时s2端口1流量满载时下发的流表 h3_to_s2_2 ='{"flow": [{"id": "2","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.3/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "2"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "1","table_id": "0"}]}' mh3_to_s2_2 ='{"flow": [{"id": "2","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.3/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "2"},"order": "0"}]}}]},'\ '"priority": "100","cookie": "1","table_id": "0"}]}' #s3流表 s3_1='{"flow": [{"id": "0","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.3/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "1"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "1","table_id": "0"}]}' s3_2='{"flow": [{"id": "1","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.1/32","ipv4-destination": "10.0.0.3/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "2"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "2","table_id": "0"}]}' #s1流表 s1_2To1='{"flow": [{"id": "0","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.2/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "1"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "1","table_id": "0"}]}' s1_3To1='{"flow": [{"id": "1","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.3/32","ipv4-destination": "10.0.0.1/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "1"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "1","table_id": "0"}]}' #h1工做时s1端口2流量空闲时下发的流表 h1_to_s1_2 ='{"flow": [{"id": "2","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.1/32","ipv4-destination": "10.0.0.3/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "2"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "2","table_id": "0"}]}' mh1_to_s1_2 ='{"flow": [{"id": "3","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.1/32","ipv4-destination": "10.0.0.3/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "2"},"order": "0"}]}}]},'\ '"priority": "100","cookie": "3","table_id": "0"}]}' #h1工做时s1端口2流量满载时下发的流表 h1_to_s1_3 ='{"flow": [{"id": "2","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.1/32","ipv4-destination": "10.0.0.3/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "3"},"order": "0"}]}}]},'\ '"priority": "101","cookie": "2","table_id": "0"}]}' mh1_to_s1_3 ='{"flow": [{"id": "3","match": {"ethernet-match":'\ '{"ethernet-type": {"type": "2048"}},'\ '"ipv4-source":"10.0.0.1/32","ipv4-destination": "10.0.0.3/32"},'\ '"instructions": {"instruction": [{"order": "0",'\ '"apply-actions": {"action": [{"output-action": {'\ '"output-node-connector": "3"},"order": "0"}]}}]},'\ '"priority": "100","cookie": "3","table_id": "0"}]}' headers = {'Content-type': 'application/json'} num=0 #下发流表,地址由ODL上得到 #下发s1与s3的流表 response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/0', body=s1_2To1, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/1', body=s1_3To1, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:3/flow-node-inventory:table/0/flow/0', body=s3_1, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:3/flow-node-inventory:table/0/flow/0', body=s3_2, method='PUT',headers=headers) #s2调用h2到1的流表 response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:2/flow-node-inventory:table/0/flow/0', body=h2_to_s2_1, method='PUT',headers=headers) while num < 4 : s1_uri = 'http://127.0.0.1:8181/restconf/operational/opendaylight-inventory:nodes/node/openflow:1/node-connector/openflow:1:2' s2_uri = 'http://127.0.0.1:8181/restconf/operational/opendaylight-inventory:nodes/node/openflow:2/node-connector/openflow:2:1' #获取s1端口2的流量 response, content = http.request(uri=s1_uri, method='GET') content = json.loads(content) statistics = content['node-connector'][0]['opendaylight-port-statistics:flow-capable-node-connector-statistics'] s1_bytes1 = statistics['bytes']['transmitted'] #0.5秒后再次获取 time.sleep(0.5) response, content = http.request(uri=s1_uri, method='GET') content = json.loads(content) statistics = content['node-connector'][0]['opendaylight-port-statistics:flow-capable-node-connector-statistics'] s1_bytes2 = statistics['bytes']['transmitted'] s1_speed=float(s1_bytes2-s1_bytes1)/0.5 if s1_speed !=0 :#获取有效的速度 print ('s1端口2速度:') print s1_speed #在检测到s1端口2流量空闲时发的流表 if s1_speed < 1000 : print(' s1端口2空闲,h1数据包改成往s1端口2经过') response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/2', body=h1_to_s1_2, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/3', body=mh1_to_s1_2, method='PUT',headers=headers) #在检测到s1端口2流量满载时发的流表 else : print(' s1端口2满载,h1数据包改成往s1端口3经过') response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/2', body=h1_to_s1_3, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/3', body=mh1_to_s1_3, method='PUT',headers=headers) #获取s2端口1的流量 response, content = http.request(uri=s2_uri, method='GET') content = json.loads(content) statistics = content['node-connector'][0]['opendaylight-port-statistics:flow-capable-node-connector-statistics'] s2_bytes1 = statistics['bytes']['transmitted'] #0.5秒后再次获取 time.sleep(0.5) response, content = http.request(uri=s2_uri, method='GET') content = json.loads(content) statistics = content['node-connector'][0]['opendaylight-port-statistics:flow-capable-node-connector-statistics'] s2_bytes2 = statistics['bytes']['transmitted'] s2_speed=float(s2_bytes2-s2_bytes1)/0.5 if s2_speed !=0 :#获取有效的速度 print ('s2端口1速度:') print s2_speed #在检测到s2端口1流量空闲时发的流表 if s2_speed < 1000 : print(' s2端口1空闲,h3数据包改成往s2端口1经过') response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:2/flow-node-inventory:table/0/flow/1', body=h3_to_s2_1, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:2/flow-node-inventory:table/0/flow/2', body=mh3_to_s2_2, method='PUT',headers=headers) #在检测到s2端口1流量满载时发的流表 else : print(' s2端口1满载,h3数据包改成往s2端口2经过') response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:2/flow-node-inventory:table/0/flow/1', body=mh3_to_s2_1, method='PUT',headers=headers) response, content = http.request(uri='http://127.0.0.1:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:2/flow-node-inventory:table/0/flow/2', body=h3_to_s2_2, method='PUT',headers=headers) odl = OdlUtil('127.0.0.1', '8181') odl.install_flow()
二.实验场景
服务器h2 h3上各自有不一样的服务,h1是客户端。实现一个负载均衡的北向程序,当h2和h3向h1传输数据时,北向应用根据链路的使用情况动态的调整路由规则。
主要思路:过必定的时间后,下发新的流表,使得流量流向另外一个交换机,从而实现负载均衡。相似于轮询法。python
三.视频演示
1.打开ODL
编程
2.查看拓扑
json
3.查看拥塞情况
ubuntu
4.调用负载均衡
服务器
5.负载均衡调用后的结果
cookie
三.小组分工网络
多级流表将数据包的处理逻辑划分为多个子逻辑,并由多张流表分别来匹配和处理,从而使得数据包的处理变成了一条流水线。单流表是没法知足复杂的业务逻辑要求的。多线程
多级流表的设计使得流表项聚合成为可能,节省了流表空间,可减小流表项个数,也提升了编程处理逻辑的灵活性。架构