EoIP Tunnel speed issue

RouterOS Wireless Networking
1

Hello.
On attached picture you can see device connections.

The issue is:
When I made speed test from device #5 or #6 to #7 the speed is around 200Mbps.
When speed test has been made from device #3 or #4 to #7 - the speed is around 400Mbps.
I expect the speed on first test to be like second one.
Do you have any suggestions why the difference between this two tests is so big?

SXTsq and hAP ac configs are attached.

Thanks in advance.

# dec/25/2018 23:23:19 by RouterOS 6.43.8
# software id = KKAN-EDHD
#
# model = RouterBOARD SXTsq G-5acD
/interface ethernet
set [ find default-name=ether1 ] name=eth speed=\
    100Mbps
/interface bridge
add name=LAN
/interface eoip
add !keepalive local-address=1.1.1.2 name=\
    eoip-B remote-address=1.1.1.1 tunnel-id=2
/interface wireless security-profiles
set [ find default=yes ] supplicant-identity=MikroTik
add authentication-types=wpa-psk,wpa2-psk eap-methods="" \
    management-protection=allowed mode=dynamic-keys name=pass \
    supplicant-identity="" wpa-pre-shared-key=password wpa2-pre-shared-key=\
    password
/interface wireless
set [ find default-name=wlan1 ] band=5ghz-onlyac channel-width=\
    20/40/80mhz-Ceee country=bulgaria disabled=\
    no frequency=auto frequency-mode=regulatory-domain name=wlan \
    nv2-preshared-key=@racom_bridge nv2-security=enabled radio-name=B \
    security-profile=pass ssid=bridge wireless-protocol=802.11
/interface bridge port
add bridge=LAN interface=eth
add bridge=LAN interface=eoip-B
/ip address
add address=1.1.1.2/29 interface=wlan network=1.1.1.0
add address=192.168.0.3/23 interface=eth network=192.168.0.0
/ip dns
set servers=8.8.8.8,8.8.4.4
/ip route
add distance=1 gateway=192.168.0.1
/system identity
set name="SXTsq B"
/system routerboard settings
set auto-upgrade=yes
/tool romon
set enabled=yes id=00:00:00:00:00:03









# dec/25/2018 23:24:03 by RouterOS 6.43.8
# software id = FFH3-EPLY
#
# model = RouterBOARD 962UiGS-5HacT2HnT
/interface ethernet
set [ find default-name=ether1 ] name=eth1 speed=\
    100Mbps
set [ find default-name=ether2 ] name=eth2 speed=100Mbps
set [ find default-name=ether3 ] name=eth3 speed=100Mbps
set [ find default-name=ether4 ] name=eth4 speed=100Mbps
set [ find default-name=ether5 ] \
    name=eth5 rx-flow-control=auto speed=100Mbps tx-flow-control=auto
set [ find default-name=sfp1 ] advertise=\
    10M-half,10M-full,100M-half,100M-full,1000M-half,1000M-full disabled=yes \
    name=sfp
/interface wireless
# managed by CAPsMAN
# channel: 2462/20-eC/gn(30dBm), SSID: ssid, CAPsMAN forwarding
set [ find default-name=wlan1 ] band=2ghz-b/g/n channel-width=20/40mhz-Ce \
    distance=indoors frequency=auto mode=ap-bridge name="wlan 2.4" ssid=\
    MikroTik-000005 wireless-protocol=802.11
# managed by CAPsMAN
# channel: 5785/20-eeCe/ac(30dBm), SSID: ssid, CAPsMAN forwarding
set [ find default-name=wlan2 ] band=5ghz-a/n/ac channel-width=\
    20/40/80mhz-Ceee distance=indoors frequency=auto mode=ap-bridge name=\
    "wlan 5" ssid=MikroTik-000005 wireless-protocol=802.11
/interface wireless security-profiles
set [ find default=yes ] supplicant-identity=MikroTik
/interface bridge port
add bridge=LAN interface=eth2
add bridge=LAN interface=eth1
add bridge=LAN interface=eth3
add bridge=LAN interface=eth4
add bridge=LAN interface=eth5
/ip neighbor discovery-settings
set discover-interface-list=discover
/interface list member
add interface=eth2 list=discover
add interface=eth3 list=discover
add interface=eth4 list=discover
add interface=eth5 list=discover
add interface=sfp list=discover
add interface="wlan 2.4" list=discover
add interface="wlan 5" list=discover
add interface=LAN list=discover
add interface=LAN list=mactel
add interface=LAN list=mac-winbox
/interface wireless cap
# 
set bridge=LAN caps-man-certificate-common-names=CAPsMAN- \
    certificate=CAP- discovery-interfaces=LAN enabled=yes \
    interfaces="wlan 2.4,wlan 5" lock-to-caps-man=yes
/ip address
add address=192.168.0.5/23 interface=LAN network=192.168.0.0
/ip dns
set allow-remote-requests=yes servers=192.168.0.1
/ip dns static
add address=192.168.88.1 name=router
/ip firewall filter
add action=accept chain=input comment="defconf: accept ICMP" protocol=icmp
add action=accept chain=input comment="defconf: accept established,related" \
    connection-state=established,related
add action=fasttrack-connection chain=forward comment="defconf: fasttrack" \
    connection-state=established,related
add action=accept chain=forward comment="defconf: accept established,related" \
    connection-state=established,related
/ip route
add distance=1 gateway=192.168.0.1
/system identity
set name="hAP ac B"
/system routerboard settings
set auto-upgrade=yes
/tool graphing
set store-every=hour
/tool mac-server
set allowed-interface-list=mactel
/tool mac-server mac-winbox
set allowed-interface-list=mac-winbox
/tool romon
set enabled=yes id=00:00:00:00:00:05

mt.png

2

Did you try to observe CPU usage on devices 3 or 4 while performing tests? What does profiling show? It is known that running traffic tests on RB devices is not prefered method as it is CPU intensive and some devices show low results because of that.

3

You are right. When I test from device #5 to #7 I see that SXTsq #3 use only one of all four cores for networking. But the ussage is 73.5% not 100%. Screenshots attached.
Is this normal? I hope not.

Wireless configuration is Point to MultiPoint with EoIP tunnels. I hope the tunnels are not the reason for that.

I use UDP protocol for tests, because less cpu ussage. On TCP the speeds are lower.
hAPs 5 and 7.png
sxtsq.png

4

What you see is quite usual behaviour. Even if test could max out CPU, the result would still be quite lower than expected 400Mbps.

The lesson learned by many is: never run btest on low-CPU-power RB device as it won’t benchmark the link, it will benchmark device’s CPU. For you it was easy to spot that as you could perform similar test over same bottle neck using different test devices and you got better results. I wouldn’t drop dead if you could achieve even better throughput if you tested it using PCs running iperf.

5

Thanks for detailed information. I will try asap with iperf.

BR,

6

While testing using iperf, try different scenarios. When I tested my setup, I found out that WiFi on my device didn’t like multiple parallel streams (i.e. more than 4) in downlink (device’s CPU just maxed up even with fairly low total throughput) whereas I couldn’t get anything near wire speed on the same device if I didn’t use many (at least 8) parallel streams while testing over gig ethernet (this one could be due to testing PC OS - windows 7 … linux did not have similar problem).

7

Hello again :slight_smile:

I tried with iperf3, but results are the same. I notice that when traffic pass through SXT SA5 ac the cpu ussage on this device is 100%. Also, cpu ussage on SXTsq 5 ac is 76% but only on one core. Would not the load be distributed among all cores, not just one of them?

Profile screenshots from both devices attached.
SXT SA5 ac.png
SXTsq 5 ac.png
After these results, I updated the version of the software to 6.44 and I saw a little improvement. The speed is around 220Mbps.

8

Hello mates.
Problem has been found.

I remove EoIP configuration (now its working only with bridge) and speed is around 500 - 550 Mbps.