Score:0

MPLS Implementation OpenFlow Iperf Testing

ye flag

I have a mininet topology that consists of two hosts and two ovs switches. h1-eth0 is connected to s1-eth1 and h2-eth0 is connected to s2-eth2 and s1-eth2 and s2-eth1 is also connected. h1 is for client and h2 is for server. In flows, ARP packets are flooded. In s1, outgoing h1 packets are pushed with mpls label (i.e.55), and in s2, mpls labels popped and directed to h2. h2 outgoing packets does not a mpls implementation, they transported as ip packets. The flow tables are like this:

s1 Flow Table:

cookie=0x0, duration=0, table=0, n_packets=0, n_bytes=0, priority=1, ip, in_port="s1-eth1" actions=push_mpls:0x08847,set_field:55->mpls_label,output:"s1-eth2"

cookie=0x0, duration=0, table=0, n_packets=0, n_bytes=0, priority=1, ip, in_port="s1-eth2" actions=output:"s1-eth1"

cookie=0x0, duration=0, table=0, n_packets=0, n_bytes=0, priority=0, arp actions=FLOOD

s2 Flow Table:

cookie=0x0, duration=0, table=0, n_packets=0, n_bytes=0, priority=1, ip, in_port="s2-eth2" actions=output:"s2-eth1"

cookie=0x0, duration=0, table=0, n_packets=0, n_bytes=0, priority=1, mpls, in_port="s2-eth1",mpls_label=55,mpls_bos=1 actions=pop_mpls:0x0800,output:"s2-eth2"

cookie=0x0, duration=0, table=0, n_packets=0, n_bytes=0, priority=0, arp actions=FLOOD

When I test this testbed with iperf (h2 is client and h1 is server), I get normal bandwidth (i.e. 20GBit). But, when I use h2 as server and h1 is client, I get very limited bandwidth (such as 400 KBit). I really want to know what causes this, but I have no idea.

Appreciate for help.

I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.