Constant high latency and packet loss

[kpanchev]

Hi, I have recently noticed a high latency on my connection, including a worrying amount of packet loss. The speed tests are all good, tested with multiple tools and speed is as expected 350/35.

My BQM monitor though indicates a constant latency issues, although occasionally I can see the issue goes away, which leads me to believe my setup is not the problem.

This is what I normally get (this is from today, but all days for the past couple of months look the same with a few exceptions):

And this is the occasional improvement:

My Hub4 is set to modem mode, with a pfsense router behind it. I have tested with a wired machine direct to the Hub and the results are the same.

I have also tested for any latency that might have been introduced by the router, but the latency between the router and hub is around 1ms, so this is not an issue.

Here are my modem levels, it all looks normal to me.

3.0 Downstream channels

Channel Frequency (Hz) Power (dBmV) SNR (dB) Modulation Channel ID

36	419000000	0.8	38.6	QAM256	36
8	195000000	2	38.6	QAM256	8
6	179000000	1.9	37.6	QAM256	6
7	187000000	1.9	37.6	QAM256	7
9	203000000	1.5	37.4	QAM256	9
10	211000000	1.3	38.6	QAM256	10
11	219000000	1.2	38.6	QAM256	11
12	227000000	0.8	37.4	QAM256	12
13	235000000	0.6	37.6	QAM256	13
14	243000000	1.1	39	QAM256	14
15	251000000	1.3	37.6	QAM256	15
16	259000000	1.2	37.4	QAM256	16
17	267000000	1.8	38.6	QAM256	17
18	275000000	1.7	37.6	QAM256	18
19	283000000	1.3	38.6	QAM256	19
20	291000000	1.1	37.6	QAM256	20
21	299000000	1.5	39	QAM256	21
22	307000000	1	38.6	QAM256	22
23	315000000	1.7	38.6	QAM256	23
24	323000000	1.1	37.6	QAM256	24
25	331000000	1.3	38.6	QAM256	25
26	339000000	1.7	38.6	QAM256	26
27	347000000	1.4	39	QAM256	27
28	355000000	1.1	38.6	QAM256	28
29	363000000	1.4	39	QAM256	29
30	371000000	0.5	37.6	QAM256	30
31	379000000	0.9	37.6	QAM256	31
32	387000000	1.3	38.6	QAM256	32
33	395000000	0.8	38.6	QAM256	33
34	403000000	1.5	39	QAM256	34
35	411000000	1.3	38.6	QAM256	35

3.0 Downstream channels

Channel Lock Status RxMER (dB) Pre RS Errors Post RS Errors

36	Locked	38.605377	0	0
8	Locked	38.605377	0	0
6	Locked	37.636276	0	0
7	Locked	37.636276	0	0
9	Locked	37.355988	0	0
10	Locked	38.605377	0	0
11	Locked	38.605377	0	0
12	Locked	37.355988	0	0
13	Locked	37.636276	0	0
14	Locked	38.983261	0	0
15	Locked	37.636276	0	0
16	Locked	37.355988	0	0
17	Locked	38.605377	0	0
18	Locked	37.636276	0	0
19	Locked	38.605377	0	0
20	Locked	37.636276	0	0
21	Locked	38.983261	0	0
22	Locked	38.605377	0	0
23	Locked	38.605377	0	0
24	Locked	37.636276	0	0
25	Locked	38.605377	0	0
26	Locked	38.605377	0	0
27	Locked	38.983261	0	0
28	Locked	38.605377	0	0
29	Locked	38.983261	0	0
30	Locked	37.636276	0	0
31	Locked	37.636276	0	0
32	Locked	38.605377	0	0
33	Locked	38.605377	0	0
34	Locked	38.983261	0	0
35	Locked	38.605377	0	0

[legacy1]

Does a ping -n 200 194.168.4.100 show packet loss when you have packet loss in BQM?

can you test without pfsense (PC with the same MAC to hub in modem mode) when you get packet loss?

what the upstream show in hub?

[kpanchev]

3.0 Upstream channels

Channel Frequency (Hz) Power (dBmV) Symbol Rate (ksps) Modulation Channel ID

1	49600000	44.5	5120 KSym/sec	QAM64	1
2	43100000	44.3	5120 KSym/sec	QAM64	2
3	36600000	44.3	5120 KSym/sec	QAM64	3
4	23600000	44.3	5120 KSym/sec	QAM64	5
5	30100000	44.3	5120 KSym/sec	QAM64	4

3.0 Upstream channels

Channel Channel Type T1 Timeouts T2 Timeouts T3 Timeouts T4 Timeouts

1	ATDMA	0	0	0	0
2	ATDMA	0	0	0	0
3	ATDMA	0	0	0	0
4	ATDMA	0	0	0	0
5	ATDMA	0	0	0	0

3.1 Upstream channels

Channel Channel Width (MHz) Power (dBmV) FFT Type Modulation

7

10.4

40.8

2K

QAM128

3.1 Upstream channels

Channel Channel Type Number of Active Subcarriers First Active Subcarrier (Hz) T3 Timeouts T4 Timeouts

7

OFDMA

208

53.4

2

0

There is no observable packet loss in ping, the packet loss on BQM is ore or less constant and varying at 2-3% (difficult to say exactly as the BQM graph is not that detailed)

I have done a test with a directly connected device (wired laptop) to the hub and the results are the same.

It seems that testing from within the inside of my network does not show any observable issue, apart from ping latency varying between 10 and 100 ms.

I am starting to think BQM is using some very advanced testing techniques, or their remote site is crap, but then they are the most reputable outside testing source...

[legacy1]

You could setup BQM on PC without router to test and do a hour without upload stream and with using OBS to see what that does.

[Carley_S]

Hi @kpanchev 

Welcome back to the community forums 

Sorry to hear that you're having service concerns. 

I have not been able to locate you on our systems to check at our side. 

Can you try the support kindly provided by our community member @legacy1 to see if this can shed some light on the issue? 

Please let us know if you are having any further issues or have any questions at all. We'll be here to help on the community forums if needed :). 

[IPFreely]

---

@kpanchev wrote:

I am starting to think BQM is using some very advanced testing techniques, or their remote site is crap, but then they are the most reputable outside testing source...

---

Nope, they're pinging you once a second. That ping is being answered by your pfSense. The BQM looks consistent with heavy usage of either your connection itself or the resources on the pfSense box. May I suggest testing latency from your pfSense box to the outside world, I believe it can monitor it, and if that's okay putting it down to the pfSense being slow to respond to BQM.

If latency and loss are normal in regular usage it's fine. Your pfSense box looks underpowered at first impression.

[kpanchev]

So, I have an update to this. Last night I was monitoring my outgoing latency and packet loss and everything was within the expected 10 to 15 ms and also no packet loss.

This is the result from the last hour with monitoring IP 194.168.4.100 (VMs own DNS server)

Then I had another look at the BQM monitor and decided to eliminate any DNS possible issues by using my IP address instead of the domain name, and here are the results:

The initial irregularities are due to me trying to saturate my link by running continuous iperf tests, which showed consistent up/down speeds of 36/360 as expected on my connection.

I will look into the DNS resolution issue further, but this is not an issue with the VM connection, but rather with the DNS server the BQM monitoring system might be using.

[kpanchev]

Just tested it with a different DDNS service and the latency jumped straight up, same with packet loss.

Checking further on thinkbroadband's forums, it seems I am not the only one seeing this issue with FQDN vs IP monitoring.

[legacy1]

My BQM is by FQDN no-ip and its fine

[Carley_S]

Thanks for keeping us and the forums updated @kpanchev 

Please let us know if you are having any further issues or have any questions at all. We'll be here to help on the community forums if needed :).