Hi all, so now with the latest epmp AX firmware release we are looking to pop a couple up, swap a couple customers to AX SMs and leave the rest which are Force 300-25. Swapping from a 3K to a 4K, keeping it at 40mhz wide channel, would we be even gaining much?
Hi. You can expect this upgrade will be very similar to what happened when the ePMP 3000 came out, and everyone had 1000/2000 series SM’s out there.
If you remember at that time, simply upgrading the AP from a 1000 or 2000 to an ePMP 3000 and continuing to run all Force 200’s gave some slight improvements because of 3000’s greater CPU and Memory, but if all your clients were still connecting at MCS7 = 64QAM because they were all 11n generation, then that was still the overall limiting factor.
However, if you upgraded a few of the heaviest talkers (your top ~20% highest users) to Force 300’s, and if those connected at MCS9 = 256QAM, that could give those heavy users a bit boost, and also get them on & off the air more quickly, freeing up more AP resources for everyone else at a result.
So, it’ll be similar now – if all you do is just plop in an ePMP 4500 AP, there may be some slight differences and benefits felt, but if everyone remains as a 300/3000 11ac generation client, then they all will have MCS9 = 256QAM modulations, and ultimately will top out at about the same throughput.
BUT YES – if you upgrade some of those clients to 4000/400 11ax generation SMs – and if the clients you upgrade first are your top users, your heaviest users, AND if you get signals good enough to get those F400 series SMs into MCS11 = 1024QAM modulations, then yes, that’ll get those SMs services more quickly, and should free up airtime for the rest of the F300 clients, which could mean an overall benefit to everyone on a heavily loaded AP.
Sounds about right. Any noticeable downsides to the AP upgrade? The 3k AP upgrade caused all sorts of latency increases and upload throughput reductions for our customers. Cambium didn’t see a problem until they did (many months later) before it was fixed. I lost a handful of customers over that, which won’t be repeated.
Appreciate the insight. I will have to see if anyone can comment on a full AX deployment, both AP and SM and see how much more is really gained going from AC to AX. Thank you!
@Eric_Ozrelic dont you have an all AX sector?
ATM, there are some temporary limitations that are holding back the 4500 from greatness. As an example, the 4500 in TDD mode does not have 8x8 MU-MIMO or beamforming enabled… it’s effectively performing like a 2x2 radio. In addition, 80MHz channel width on e4k 5GHz still needs optimization. Lastly, the TDD scheduler still needs more optimization to reduce latency and jitter. The current net result is that the 4500 can perform slightly better or a little worse then a 2x2 e3k AP under different load and RF conditions.
@Eric_Ozrelic thank you for your info! When would the listed limitations be rectified? Is this in the works or is this a further down the road situation?
It has been a little bit, do you know if the 8x8 TDD has been fixed? We are looking at replacing some 3k APs
Be on the lookout for the upcoming 5.8 firmware beta which will fix beamforming in TDD mode.