i presume all wisp has some percentage of gamers on network, so I was wondering how is it working out for you with them , I receive constant complaining form them regarding jitter!
APs 3000 , all 4.6.1 working good , gamers custamers have good signals and no retransmission, I am lowering MCS regardless trying to minimize packet drops and capacity packet drops …but they acure regardless.
Maybe maybe it is not affecting performance that much but,
even today after 4.6.1 it looks like epmp 2000 works beter in this regards N clients do not have any packet drops ever … ( or they do not report correctly dont know ) but all my ac clients have packet drops!
When I test with mikrotik tools I dont see any losses … except crappy latency and high standard deviation!
Thank god I dont have much voip/gamers clients but I need to do something! Client has beter gaming experience using mobile tethering while almost dead signal is disappointing!
We have quite a few gamers, most casual, some serious, streaming twitch to audiences, etc. As you noted, no issues on ePMP 1000 or PMP 450/450i. When we deployed our first ePMP 3000, the first customer on it was a gamer and was literally a rocks throw from the tower and had similar latency and jitter complaints, even with the best signal possible. We did see more variation on the ePMP 300 than we see on the other platforms.
We ended up moving him over to an ePMP 2.4 AP that was on the tower which resolved the issue We had the ePMP 3000 up for evaluation have since replaced the ePMP 3000 with an ePMP 1000. The location did not justify the ePMP 3000 and an ePMP 1000 can easily handle the load. Most of our customers are at distance and we weren’t particularly impressed with the reach of the AC gear, but can’t complain about its capacity and it seems pretty solid under 4.6.1. I’ve seen other posts complaining of the jitter issue but thought it was mostly resolved with 4.6.1
VOIP also. On two towers we have had to leave 2000AP in place just for customers with VOIP and a few gamers. The CPE can be a N or AC but if I move them over to the one of the 3000 sectors , where in many cases they get much better RSSI, they immediately start complaining or their IT people start emailing pages of pings showing 100+ ms swings and the occasional dropped packet. Even on 3000 AP’s with very little load (frame time rarely above 50%)
We have exact same issue as what you all have described and we launched using all epmp 3000 units on the tower sites. We have observed that the ping/jitter remains high for our clients that are on busy sectors(which is around 50% frame time) and our not so busy sectors(10-20% frame time). We have ran ping and tracert tests to different servers and the higher ping/jitter results are from the SM to said server. Testing from the AP, Router, ptp link all have expected ping results.
I have an associate that runs a small wisp as well but uses Ubiquiti Rocket Prism AC for ptmp and their ping/jitter results on busy sectors are better than our not so busy sectors. I have seen a lot of posts on this topic before and most answers to the high ping is either high frame time percentage and the way TDD is designed to function. As per the Ubiquiti Rocket Prism AC spec sheet it uses TDMA, not sure if that is why the Ubiquiti ping/jitter SM results are better. Our deployment and my associates deployment are very similar, very rural, and the only difference is ptmp equipment choice.
Part of the issue is that 1000/2000 series gear has a choice between 5ms and 2.5ms frame sizes, with 2.5ms frames having lower latency and lower throughput. However, on the 3000 series gear, there is currently only 5ms frames… which will have higher throughput, but higher latency.
Do not worry about capacity drops unless the AP to SM link is having capacity issues. Worry about the other packet drop counters. They maybclimb a bit for several reasons, smart speed does cause this and so does a client side device which supports low power ethernet.
If the wireless packet drop counter is increasing then check the link for stability and good (not good enought) margin.
We have a lot of gamers and more than a few VOIP using clients (most are using zoom or teams) and the latency issue is moot unless they are testing to our gateway and not to some place off our network. Our service contract is very clear on this and we stick to it. We do not take responsibility for the clients network devices nor for data once it leaves our network so pages of pings showing latency issues are generally useless.
There is a difference between the e1k and the e3k AP for latency stability but we have only seen a minor 15ms swing in ping times from the tower to the client. It does make a difference if the sector has only f300 SMs on it. Not having to talk to N clients seems to be a good way to provide improvement though thisbis anecdotal as we did not do this purposely.
i didn’t see much difference regarding this problem by by having mixed or all ac clients … but I did notice that this latency/jitter probelm is not manifesting up till 15-20 SMs … So i believe it has something to do with TDM scheduler protocol … but i’m guessing