Mesh Mashup Redux - NETGEAR's Orbi Checks In

[Pablo Salvador]

When we ran our mashup tests on NETGEAR's Orbi, the winning product was very clear.

Read on SmallNetBuilder

Hi! Very complete set of tests. My only concern and doubt is regarding the difference results between the downlink and uplink throughput in the tests? Are you using a different ISP plan for uplink and downlink? Thanks for such an amazing work.

 

[thiggins]

Hi! Very complete set of tests. My only concern and doubt is regarding the difference results between the downlink and uplink throughput in the tests? Are you using a different ISP plan for uplink and downlink? Thanks for such an amazing work.

All testing is done on a private network. No ISP is involved.
There is always a difference between up and downlink traffic due to the power levels of router/AP and client device and number of transmit streams.

 

[Pablo Salvador]

ok, got it. Thanks a lot for your reply!
Still, I don't understand such a huge difference (in other analyses you performed UL and DL were very similar). Regarding the number of streams, correct me if I'm wrong, you need to have the same number of streams in the receiver and transmitter devices to take advantage of higher data transmission rates.

 

[thiggins]

Regarding the number of streams, correct me if I'm wrong, you need to have the same number of streams in the receiver and transmitter devices to take advantage of higher data transmission rates.

That's true. But you can get a range of throughput values at the same link rate depending on signal strength.

 

[Pablo Salvador]

Sure, that relationship you mention between signal strength and link rate is true, but that would imply you have very different conditions for both uplink and downlink, a very asymmetry link. In an area with several APs, the uplink and downlink interference can be different. Thus, it can lead the asymmetry of downlink and uplink signal qualities. Was that crowded the network scenario? Thx and sorry for so many questions, just wanted to understand it properly.

 

[thiggins]

Sure, that relationship you mention between signal strength and link rate is true, but that would imply you have very different conditions for both uplink and downlink, a very asymmetry link. In an area with several APs, the uplink and downlink interference can be different. Thus, it can lead the asymmetry of downlink and uplink signal qualities. Was that crowded the network scenario? Thx and sorry for so many questions, just wanted to understand it properly.

Our testing is done in a quiet RF environment with no other networks operating.

Yes, conditions are different, as is common for many networks. Downlink has a 4x4 transmitter talking to a 2x2 receiver; vice-versa for uplink. Link rate for the connection is the same. Effective link gain would be better running uplink with four receive channels in operation.

 

[Pablo Salvador]

Ok, thanks a lot for your reply! A couple comments:
- The reciprocity theorem guarantees the same propagation conditions in the uplink and downlink directions, and as your testing is done in a quiet RF environment with no other network operating, then symmetric radio conditions should be obtained (indeed you state link rate for the connection is the same)
- Regarding the number of streams, the transmitter is equipped with Nt antennas and the receiver has Nr antennas, the maximum number of antennas that will be transmitted in parallel will be the minimum between Nt and Nr, in your testing that number then would be 2. Therefore, I still don't understand such a big difference between UL and DL throughput tests.

Thanks a lot for the discussion.