bandwidth loss issue due to the nsh #poll-notice


josephjin@...
 
Edited

Hi,
    we found that the nsh make our TCP bandwidth loss. scene as following
    threre ar two host: vm1(192.168.128.2), vm2(192.168.128.3)
    1: host(vm1/vm2) send traffic to vpp_server0
    2: vpp_server0 configrate the classify table and session for river the traffic to vpp_server1
    3: vpp_server1 configrate the classify table and session to send the traffic back to vpp_server0
    4: vpp_server0 send traffic to host

    vpp_server 0/1 nsh configration (hit-next 27 mean nsh-classifier)
    vpp_server0:
    classify table mask l3 ip4 src
    classify session hit-next 27 table-index 0 match l3 ip4 src 192.168.128.2 opaque-index 256255
    classify session hit-next 27 table-index 0 match l3 ip4 src 192.168.128.3 opaque-index 256255
    set interface l2 input classify intfc pipe1000.0 ip4-table 0

    create nsh map nsp 1001 nsi 255 mapped-nsp 1001 mapped-nsi 255 nsh_action pop encap-none 3 0
    create nsh map nsp 1002 nsi 255 mapped-nsp 1002 mapped-nsi 255 nsh_action pop encap-none 3 0

    vpp_server1:
    classify table mask l3 ip4 dst
    classify session hit-next 27 table-index 0 match l3 ip4 dst 192.168.128.2 opaque-index 256511
    classify session hit-next 27 table-index 0 match l3 ip4 dst 192.168.128.3 opaque-index 256767


    we use iperf to test vm1/vm2 tcp bandwidth, it just 1Gbps
    if we just modify the configration of vpp_server1 as following, tcp bandwidth will increate to 6Gbps we expected

    classify session hit-next 27 table-index 0 match l3 ip4 dst 192.168.128.3 del
    classify session hit-next 27 table-index 0 match l3 ip4 dst 192.168.128.3 opaque-index 256511

    the vpp version we used is 19.08. we don't know why nsh cause this issue.

    B.R.
    joseph

Results

See Who Responded