Roaming Interface Congestion?

When I was recently in Austria I noticed that in the evening hours, the superb network speed I got over the LTE network in a town in the mountains at the place I was staying went from 40-50 MBit/s during the daytime to just 2-3 Mbit/s in the evening. Typical dissapointing busy hour overload at the LTE cell site due to underdimensioning I thought at first. But then I noticed that this was not quite the case.

Two things made me question my assumption:

The first thing I found a bit suspicious was that the ping round trip delay times remained at 80 ms despite doing a throughput test. This is very unusual because if the local cell site is overloaded, that value usually goes way up into the 500-2000 ms region.

Then, to cross check, I manually selected a network that was not preferred by my home network operator without changing location. And to my great surprise, the throughput to the same server was suddenly back to 40-50 Mbit/s even in the evening hours when the other network would only deliver 2-3 Mbit/s. Also, the round trip times during the throughput test went up to around 500 ms, a typical indication that the pipe is fully utilized.

So my working assumption is that this was NOT caused by the local LTE cell site or backhaul but rather by a congested roaming interface that connects the visited network to my home network operator. Further, I think the problem is on the roaming leg of the visited network operator and not of the home network operator as throughput was just fine in the second visited network.

I’m glad there is choice!

One thought on “Roaming Interface Congestion?”

  1. I would assume, that visited network operator has a policy to grant inbound roamers a worse QoE.

Comments are closed.