CRS112-8G-4S routing power far below specs (unusable)

Hardware MikroTik hardware general
1

I realized a simple test bench with CRS112-8G-4S, the results is astonishing slow.

The test bed is made in this way:

VM linux ------- CRS (192.168.100.1/24) -------- myLANrouter (10.100.1.1/24) --------- Internet

To avoid any variability related to Internet I run:
iperf3 -s on 10.100.1.192 host

The test has been done with iperf3 -c 10.100.1.192 and the result is about only 31 Mbps, 2.5kpps , 100 % CPU

Replacing the CRS with an hAP, configured in the same way, the iperf3 result is wire speed (at the hAP port), 94 Mbps
This is not just a question of CPU MHz, there is something wrong on CRS112-8G-4S model.

This is the configuration used:

/ip pool
add name=dhcp_pool1 ranges=192.168.100.2-192.168.100.254
/ip dhcp-server
add address-pool=dhcp_pool1 disabled=no interface=ether4 name=dhcp1
/ip address
add address=192.168.100.1/24 interface=ether4 network=192.168.100.0
/ip dhcp-client
add default-route-distance=0 dhcp-options=hostname,clientid disabled=no interface=ether1
/ip dhcp-server network
add address=192.168.100.0/24 gateway=192.168.100.1
/ip dns
set allow-remote-requests=yes
/ip firewall nat
add action=masquerade chain=srcnat out-interface=ether1
/system clock
set time-zone-name=Europe/Rome

Best regards

VLAN & CPU load on CRS-112-8G-4S-IN
2

Sorry to disappoint you, but according to the specs at https://routerboard.com/CRS112-8G-4S-IN this is normal for that device.

I typically look at the 25 IP filter rules, 512 byte packets to determine the real world routed throughput of a device, and that page shows the max throughput as 32Mbps.

The hAP is much more powerful, with a max throughput of ~170 Mbps.

You might be able to coax a little more performance out of the CRS by using fasttrack.

The CRS is really too underpowered to be a router for all but the slowest of Internet connections.

3

As the name implies, it is intended as a switch.

4

Most of CRS router got enough power to manage a common internet connection of max 100 Mbps (most of the time 50 or 30 FTTC)
As a switch this device will loose any market appeal. In my experience Is wrong to say that a CRS is intended as a switch, only the CRS112 is just a switch.

This way to read specs is not exact, because throughput depends from packet size. On the worst case CRS112 specs reports 8 kpps.
Anyway I’m pretty sure Iperf3 use large packets.
In my case, the maximum packet throughput is 2.5 kpps, but without any firewall filter, just a masquerade for a single client.
8 kpps would be more then I need.

Mikrotik official support neglects to explain me why the router doesn’t comply kpps specs and suggested to me buying a CCR router (?!?!?)
I wouldn’t like to swear here, I limit myself to say “this is not a serious way to deal with costumers”.
This is the second time I face problems with a new Mikrotik router model.

If this is Mikrotik approach to new product test and costumers dealing, I strongly suggests to other Mikrotik users to avoid buying new Mikrotik product without prior checking user feedback.
Clearly, in this way, nobody will take the risk to buy a new model.

5

maybe you should return it and buy something else.
I have two CRS112’s in service. both route 50mbps links just fine. I wouldn’t expect such a low power CPU to route any faster. The specs on the site show a best case. Bottom line, if you need faster routing, buy a router that can handle double what you expect to need.

6

Hi IntrusDave, do you use “masquerade”?

Thanks

7

yes

8

Well, sorry to bother you. My testbed is made just of masquerade. Please, What kpps value are you getting during a speedtest (OOKLA)?
You could get it easily:

interface monitor-traffic [interface]

Maybe my CRS112 is simply defective.