Evaluating Peplink SpeedFusion: Part 2

evaluating peplink speedfusion part 2

1 + 1 = 1?

Another question that often comes up when evaluating SpeedFusion bonding performance is that

I have 15Mbps on one 4G LTE. How come when I bond two 4G LTE, the bonded bandwidth is still 15Mbps?

There is something behind the scene.

Fixed Bandwidth Pool

First thing first. It is important to make sure that we are measuring SpeedFusion bandwidth properly. Here are steps for evaluating Peplink SpeedFusion.

Here the problem is likely to do with

4G LTE mobile carriers have a fixed pool of bandwidth at the backhaul and they share this fixed pool between their SIM cards.

4g lte mobile carrier fixed bandwidth pool

This way regardless of how many 4G LTE connections we use on the Peplink routers, the total bonded SpeedFusion bandwidth will be limited by this 15Mbps fixed bandwidth pool at the carrier backend.

And it is rather easy to verify.

Speed Test. In Parallel.

If the problem is with bandwidth sharing at the 4G LTE carrier backhaul, we should be able to reproduce it without using Peplink SpeedFusion at all.

To isolate this problem, we will go back to an old friend – speedtest.net.

Speedtest.net measures how much bandwidth each of the two 4G LTE connections has to the nearest speedtest.net server.

If we do two speed tests at the same time on the two 4G LTE connections, we will be able to see how much bandwidth the 4G LTE mobile carriers can provide to each of the two 4G LTE, when they are used concurrently.

evaluating peplink speedfusion part 2 the setup

This is our setup.

Setup #1 on the left is connected to one Pepwave MAX BR1 with one 4G LTE SIM card from KPN, a 4G LTE carrier in the Netherlands.

Setup #2 on the right is connected to another Pepwave MAX BR1 with also one 4G LTE from KPN.

The two setups are completely separate. Simply two laptops each with one 4G LTE connection via a MAX BR1. There is no Peplink SpeedFusion bandwidth bonding.

If you do not have a MAX BR1, you can insert the 4G LTE SIM cards in two different smartphones or two 4G routers of any kind and do the same test.

Now we do speed tests.

First, we do a speed test on the setup #1.

pepwave max br1 with kpn 4g lte setup #1

14Mbps of download bandwidth and 13.4 upload bandwidth to the speedtest.net server. Decent.

And then we do a speed test on setup #2.

pepwave max br1 with kpn 4g lte setup #2

10.4Mbps of download bandwidth and 15.2 upload bandwidth to the speedtest.net server. Not so much different that the other setup.

Now it gets interesting. We do the speed tests at the same time.

Imagine.

In a world where there is no bandwidth limitation and bandwidth sharing at the 4G LTE carrier backhaul, if we do the two speed tests concurrently we should expect roughly the same results on each of the 4G LTE as when we do the two speed tests separately.

The world never fails to amaze us.

Bandwidth Ceiling

These are the speed test results when the tests are done concurrently.

concurrent speed tests on pepwave max br1 setup #1 and #2

When speed tests are done concurrently, download and upload bandwidth of each of the 4G LTE drop to roughly half of what we had before.

This points to a fixed bandwidth pool of around 15Mbps download and upload at the 4G LTE carrier backhaul – shared between the two 4G LTE connections.

If we do SpeedFusion bandwidth bonding on these two 4G LTE SIM cards, with each of the 4G LTE giving roughly half of 15Mbps, SpeedFusion will bond them to a throughput of 15Mbps.

This 15Mbps bonded throughput will mislead the users to believe that SpeedFusion is not bonding the two 4G LTE connections.

In fact, if we take into account the fact that the two 4G LTE are constantly nudging each other for more bandwidth in a fixed bandwidth pool, and we take away the 10-20% bandwidth cost from packet encapsulation, user may see bonding two 4G LTE gives less bandwidth than using only one 4G LTE i.e. 1 + 1 = 0.9.

Is there a way around this?

Bonding 4G LTE

From our experience,

This bandwidth sharing problem is quite common for 4G LTE from the same mobile carriers. And sometimes we do also see this problem for 4G LTE from different mobile carriers, especially when they are sharing the backhaul infrastructure or using same cellular frequencies.

Be sure to do the parallel speed test above if you suspect that you have fallen a victim of this bandwidth sharing problem.

A number of 4G LTE mobile carriers also have 4G LTE connection offerings with dedicated bandwidth. These will be ideal for bandwidth bonding.

And keep in mind that there is a world of connectivity options other than 4G LTE that SpeedFusion can use to boost bandwidth and increase resiliency.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *