Multipath routing with Group table

ovs command

ovs-vsctl set Bridge 0e2bf920-f12d-49c3-b9fe-26d888becf77 protocols="OpenFlow13"
ovs-vsctl add-br br0
ovs-vsctl add-port br0 port1 -- set Interface port1 type=internal
ovs-vsctl add-port br0 port2 -- set Interface port2 type=internal
ovs-ofctl -O OpenFlow13 add-group br0 group_id=111,type=all,bucket=output:2
ovs-ofctl -O OpenFlow13 dump-groups br0
ovs-ofctl -O OpenFlow13 show br0
ovs-vsctl add-port br0 port1
ovs-vsctl add-port br0 port2
ovs-ofctl -O OpenFlow13 dump-ports br0
ovs-ofctl -O OpenFlow13 dump-groups br0
ovsdb-client dump
#### Modify the group
ovs-ofctl -O OpenFlow13 mod-group br0 group_id=111,type=all,bucket=output:1,2
ovs-ofctl -O OpenFlow13 add-flow br0 "table=0, actions=group:111"
ovs-ofctl -O OpenFlow13 del-flow br0 "table=0, actions=group:111"
ovs-ofctl -O OpenFlow13 del-flows br0 "table=0"
ovs-ofctl -O OpenFlow13 del-groups br0 group_id=111
ovs-ofctl -O OpenFlow13 add-flow br0 "table=0, actions=group:200"

转载于 https://www.hwchiu.com/2014-06-25-multipath-routing-with-group-table-at-mininet.html

Purpose

在Group table中，有一個類型為select，此類型的group會隨機執行底下的其中一個bucket。若我們將所有的output action都放進這個group中，則switch會將封包隨機導向不同的port，藉此達成multipath routing的功用。

 

 

Environment

使用下列的圖作為我們的網路環境，在此圖中。S1~S5都是支援OpenFlow 1.3的OpenFlow switch，左邊的Host 1則是一個Sender，會對於右邊的九個Host發送資料

Step

• 使用mininet搭配其script來創造網路拓墣，該script可以在此找到 group.py

  
  

  1	mn --custom group.py --topo group

• 讓所有的創造的openvswitch都支持openflow 1.3

  
  

  1 2 3 4 5

  ovs-vsctl set bridge s1 protocols=OpenFlow13 ovs-vsctl set bridge s2 protocols=OpenFlow13 ovs-vsctl set bridge s3 protocols=OpenFlow13 ovs-vsctl set bridge s4 protocols=OpenFlow13 ovs-vsctl set bridge s5 protocols=OpenFlow13

• 在S1上面加入一個group table，此group table能夠把封包給隨機導向Port 1,2,3。

  
  

  1	ovs-ofctl -O OpenFlow13 add-group s1 group_id=5566,type=select,bucket=output:1,bucket=output:2,bucket=output:3

• 在S1上面加入一個Flow entry，所有從Host1進來的封包，都去執行剛剛所創立的group table。

  
  

  1	ovs-ofctl -O OpenFlow13 add-flow s1 in_port=4,actions=group:5566

• 由於本實驗沒有採用任何Controller，因此要手動的寫入Flow entry到其餘的Switch。

• 在S1上面加入剩下的Flow entry，使得送回Host1的封包能夠順利抵達Host1

  
  

  1 2	ovs-ofctl -O OpenFlow13 add-flow s1 eth_type=0x0800,ip_dst=10.0.0.1,actions=output:4 ovs-ofctl -O OpenFlow13 add-flow s1 eth_type=0x0806,ip_dst=10.0.0.1,actions=output:4

• 在S2、S3、S4上各加入兩條Flow entry，讓封包能夠通過

  1 2 3 4 5 6

  ovs-ofctl -O OpenFlow13 add-flow s2 in_port=1,actions=output:2 ovs-ofctl -O OpenFlow13 add-flow s2 in_port=2,actions=output:1 ovs-ofctl -O OpenFlow13 add-flow s3 in_port=1,actions=output:2 ovs-ofctl -O OpenFlow13 add-flow s3 in_port=2,actions=output:1 ovs-ofctl -O OpenFlow13 add-flow s4 in_port=1,actions=output:2 ovs-ofctl -O OpenFlow13 add-flow s4 in_port=2,actions=output:1

• 在S5上根據destination ip來把封包導向不同的host

  
  

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

  #IP ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.2,actions=output:4 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.3,actions=output:5 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.4,actions=output:6 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.5,actions=output:7 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.6,actions=output:8 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.7,actions=output:9 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.8,actions=output:10 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.9,actions=output:11 #ARP ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.2,actions=output:4 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.3,actions=output:5 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.4,actions=output:6 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.5,actions=output:7 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.6,actions=output:8 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.7,actions=output:9 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.8,actions=output:10 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.9,actions=output:11

• 由於本實驗要觀察的是Host1送過來的封包能否走不同路徑，對於送回給Host1的封包就固定於同一條路徑(S5 - S2 - S1)

  Copy

  1 2	ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0800,ip_dst=10.0.0.1,actions=output:1 ovs-ofctl -O OpenFlow13 add-flow s5 eth_type=0x0806,ip_dst=10.0.0.1,actions=output:1

• 接下來依序執行下列指令來產生網路流量

  
  

  1 2 3 4 5 6 7 8 9

  mininet> iperfudp 1G h1 h2 mininet> iperfudp 1G h1 h3 mininet> iperfudp 1G h1 h4 mininet> iperfudp 1G h1 h5 mininet> iperfudp 1G h1 h6 mininet> iperfudp 1G h1 h7 mininet> iperfudp 1G h1 h8 mininet> iperfudp 1G h1 h9 mininet> iperfudp 1G h1 h10

• 接下來觀察每個switch的flow table。結果如圖

  
  

  1 2 3

  mininet> sh ovs-ofctl dump-flows s2 -O OpenFlow13 mininet> sh ovs-ofctl dump-flows s3 -O OpenFlow13 mininet> sh ovs-ofctl dump-flows s4 -O OpenFlow13

• 在圖中可以觀察到，S2、S3、S4上面都有流量經過，證實了S1使用了group table會將不同的flow給隨機執行不同的buckets，在此範例中則是會導向不同的port。

[root@kunpeng82 devuser]# ovs-ofctl dump-groups s1
NXST_GROUP_DESC reply (xid=0x2):
 group_id=5566,type=select,bucket=bucket_id:0,actions=output:"s1-eth1",bucket=bucket_id:1,actions=output:"s1-eth2",bucket=bucket_id:2,actions=output:"s1-eth3"
[root@kunpeng82 devuser]#