MG21E Data Usage

[RestaurantGuy]

We have set up an MG21E as a failover connection for a restaurant.  Works very well and fails over fast and none of the issues we had using the USB devices connected directly to the MX.

Our issue is that the MG seems to be using about .13Mbs of data at a fairly constant rate when we are not in a fail over state.  With one device this is not a huge issue, but we have a shared data plan and are planning to add them to all of our locations, meaning that data usage will become an issue.

Does anyone know of a way to limit the amount of data usage or better manage data while not being the primary connection?

[Inderdeep]

@RestaurantGuy : The MG21 cellular gateway uses 6 to 8 megabytes or little more  a day of data for telemetry on dashboard and connection monitoring when the unit is idle.

Any additional Cisco Meraki devices that are added to the MG21 may increase data usage further. Cisco Meraki products are cloud connected devices that report telemetry into Dashboard for network monitoring purposes.

Check page no 2, above the use cases

https://media.trustradius.com/product-downloadables/F6/X7/BB87PR4HXTPE.pdf

[RestaurantGuy]

Sounds like I need to go and add the ability to send it's telemetry out through the Ethernet port to the wish list!

Thanks for the clarification.

[ww]

In case you use mx auto vpn, you can choose to not run active vpn on wan2 . 

[RestaurantGuy]

I don't see the option of controlling the secondary WAN usage of the VPN.  I thought it only created a connection to the active WAN interface?

[ww · Accepted solution]

You can check it at you mx

Sd-wan & traffic shaping > Active-Active AutoVPN

[PhilipDAth]

From my experience, Meraki devices generate about 100MB of telemetry data per month.

So if you had an MG21E with an MX plugged into it for backup, I would expect to see 200MB of usage per month if the link is never used for actual data.

[Daniel-GTVR]

Good day, I have the same problem but my MG21E is using 35 MB per day, so a total 1GB of data per month, my cellular basic package is 100 MB so I get extra fee every month event when no actual data is used,

I'm fine with de 100 Mb per month for the monitoring but what about the other 900 Mb, where would that come from ?

The MX connected behind has no VPN (single site) and traffic shapping Load Balancing is set to disable

From what I'm seing in the portal there is a constant about 3 Kbps banwith usage 24/7

Let me know if you have any clue how to lower this usage

Thanks in advance

Daniel

[Bruce]

Meraki states in this document, https://documentation.meraki.com/MG/MG_Datasheets/MG21%2F%2F21E_Datasheet,  about 6 to 8 Megabytes per day, plus you then need to add in the telemetry from the MX (which it send on all WAN ports regardless of which is active). Even so, you shouldn’t be getting to 1GB of traffic over a month.

You might need to resort to some packet captures to see exactly where the traffic is coming from and going to.

[Daniel-GTVR]

Thanks for the fallow and the idear packet caputure

I did check packet capture

thie only thing going throu right now are ping to 8.8.8.8 and response but there are a lot of them (every second about 2 or 3 ping ) I have put some result bellow

25.135.225.168 is the ip assing by the cell phone provider

But what is weird I think is the IP 192.48.236.3  This IP is refer to by ARIN as own by mobup.com whitch is a Twitter service of mobile advertising

and the IP 198.246.30.1 is the WAN1 gateway IP of the MX behind the MG21E So I guest that this is what you were saying about the MX adding bandwith on all WAN ports

I there anyway to block answering to ping on the MG21E, at least that could help for the Twitter service

Thanks

Daniel

21:22:27.482620 IP 8.8.8.8 > 25.135.225.168: ICMP echo reply, id 54227, seq 282, length 64
21:22:27.498597 IP 198.246.30.1 > 25.135.225.168: ICMP echo reply, id 54483, seq 282, length 64
21:22:27.524590 IP 192.48.236.3.443 > 25.135.225.168.50689: Flags [F.], seq 1339843958, ack 482278337, win 271, options [nop,nop,TS val 335117731 ecr 81616114], length 0
21:22:27.544577 IP 192.48.236.3.443 > 25.135.225.168.51020: Flags [F.], seq 4001798523, ack 846653120, win 271, options [nop,nop,TS val 2747332167 ecr 81616094], length 0
21:22:27.842600 IP 192.48.236.3.443 > 25.135.225.168.50689: Flags [FP.], seq 4294967265:0, ack 1, win 271, options [nop,nop,TS val 335118035 ecr 81616114], length 31
21:22:28.260598 IP 192.48.236.3.443 > 25.135.225.168.51020: Flags [FP.], seq 4294967265:0, ack 1, win 271, options [nop,nop,TS val 2747332871 ecr 81616094], length 31
21:22:28.420693 IP 25.135.225.168 > 8.8.8.8: ICMP echo request, id 54227, seq 283, length 64
21:22:28.420705 IP 25.135.225.168 > 198.246.30.1: ICMP echo request, id 54483, seq 283, length 64
21:22:28.460660 IP 192.48.236.3.443 > 25.135.225.168.50689: Flags [FP.], seq 4294967265:0, ack 1, win 271, options [nop,nop,TS val 335118668 ecr 81616114], length 31
21:22:28.482682 IP 8.8.8.8 > 25.135.225.168: ICMP echo reply, id 54227, seq 283, length 64
21:22:28.500563 IP 198.246.30.1 > 25.135.225.168: ICMP echo reply, id 54483, seq 283, length 64
21:22:29.424395 IP 25.135.225.168 > 8.8.8.8: ICMP echo request, id 54227, seq 284, length 64
21:22:29.424409 IP 25.135.225.168 > 198.246.30.1: ICMP echo request, id 54483, seq 284, length 64
21:22:29.522695 IP 8.8.8.8 > 25.135.225.168: ICMP echo reply, id 54227, seq 284, length 64
21:22:29.538561 IP 198.246.30.1 > 25.135.225.168: ICMP echo reply, id 54483, seq 284, length 64
21:22:29.656689 IP 192.48.236.3.443 > 25.135.225.168.51020: Flags [FP.], seq 4294967265:0, ack 1, win 271, options [nop,nop,TS val 2747334279 ecr 81616094], length 31
21:22:29.678548 IP 192.48.236.3.443 > 25.135.225.168.50689: Flags [FP.], seq 4294967265:0, ack 1, win 271, options [nop,nop,TS val 335119884 ecr 81616114], length 31
21:22:30.471500 IP 25.135.225.168 > 8.8.8.8: ICMP echo request, id 54227, seq 285, length 64
21:22:30.471515 IP 25.135.225.168 > 198.246.30.1: ICMP echo request, id 54483, seq 285, length 64

[Bruce]

The ICMP messages that you are seeing are part of the MX connection monitoring that it performs, https://documentation.meraki.com/MX/Firewall_and_Traffic_Shaping/Connection_Monitoring_for_WAN_Failo.... It’s also worth noting the point about the traffic flows and how they may stay pinned if you’ve had a recent failover to the WAN interface connected to the MG21.

You may want to contact support and see if there is any tuning they can perform on the WAN interface connected to the MG21 to reduce the polling interval for the connection monitoring. Something else they can do, which you may want to consider, is assign the Cellular Failover firewall rules to WAN2 on the MX so that you can use more restrictive rules for the traffic goes out the interface connected to the MG21, as opposed to the primary path.