Forum Discussion
Using TestMy.net to compare your speed to others on your beam
I was poking around the 'TestMy.net Database' option under 'DB' on the TestMy.net page, and I noticed something that might be useful for knowing how your tests compare with other users of HughesNe...
Good points all.
I was just looking for the case where all descent tests by all users associated with a city scored well, as that implies no problem within some recent time period.
Considering two threads (titles in bold below are the threads):
Slow speed
In looking at a city for the user, I could see that all test results were bad. Of course a huge number of the tests used a test size that was far too small, and thus would give a bad score because the latency "put a thumb" on the scale so to speak. Nothing here to say what the problem is, or if there really is a problem -- but it does not say it is good either. The engineers did something and performance improved for multiple customers.
Slow internet speed
In looking at a city for this user, I saw other users with good scores at the same time this user had low scores. Of course it could be a J1 beam vs J2 beam thing, but the results hinted that this is not an issue everyone is facing. In the end, it appears that the issue was that "multithreading" was turned on. In looking in the history, the users tests were good before this feature was turned on, so the multithreading test may have been the issue.
Perhaps a Big Chungus was lodged in the Level 3 pipes between one of the servers used for the multithread test and the gateway causing the whole test score to look bad. :-)
Recently, I have been seeing really good speeds. In fact, I was thinking they were unusually good. Of course, I mostly just watch the latency as it is an unobtrusive way to have an idea of how busy things are without running a testmy -- kind of like guessing at how many people are in the store by tracking your average wait time at a grocery checkout.
And things have seemed empty, so I figured I would run a bunch of testmy in prime time which I don't normally do because I don't like the idea of wasting bandwidth when it is most in demand. The results look like they came from 4am -- this beam feels empty, like right after I first was installed in 2017, and again right after that first gateway upgrade in 2017 that doubled the IPGW that were showing up. I had other testmy result recently that were very good, but as they were run in less busy times, I figured it was not that unusual. In looking at TestMy scores for cities that I have a been associated with, I see a bunch of bad scores that have dragged the averages down on the recent tests list. Of course, many of these used the wrong test size but maybe TestMy backed the test size down since they were around 1Mbps?
There is a fellow reporting speed issues, and in his early tests, I noticed he was associated with a city at the southern edge of my beam according to SatBeams, but I figured that SatBeams is probably off a bit for the beam placement, and the location data is probably only crudely lining up with the beam, so he could be on a beam to the south. In more recent tests, the location data gave a city at the northern edge of my beam, so unless this location data is crazy off at times, it seems there is a fair chance he is on the same beam as me.
Also, I noticed his tests show remarkably consistent speed throughout the test, usually considered to be a sign of low congestion.
If he is in my beam, could other people on this beam be seeing the same thing he is? If so, could the high speeds I am seeing be the result of a group of other users on the beam being slowed way down by something? Could the rash of new testmy results with low scores at cities I have been associated with in the past be related?
I am showing J2CHY082HNSIGW1204 and have had this since 10/11 with OutRoute 4. I noticed my good speed some days later about the time the fellow posted about his low speed. I have seen in the past cases on this site where users report consistently bad speeds, then something happens and they are restored. Could a gateway server that handles a group of customers on the beam just need a reboot or something?
For you long time satellite users that have seen this before, was the resolution something along these lines? On the flip side, have you ever had a period of a week or so where the speed is insanely good?
What do you think? Of course, there is a lot of guessing here, but with some communication some light might be shed on what is going on.
I've had this 'conspiracy theory' that whenever the beam takes a sharp decline in performance, there's possibly always one (or two) recently added people to the beam that's bringing it down for everyone. One possibility is the person doesn't realize they're causing their own problem and discontinues, so the beam frees up.
Another possibility is the person was new, in their initial 20-day period, and didn't realize they were consistently dragging it down until they finally hit the Fair Access Policy. Whatever was causing the problem for everyone was therefore relieved.
If the problem was consistent enough, it was likely something like a constantly running HD security camera - usually something that takes a lot of resources and the person wouldn't even think twice about until it finally no longer worked.
Of course this is all supposition on my part, but have seen it happen often enough. Performance comes and goes. To be honest, it's probably a lot simpler explanation like the network periodically adjusting the load.To be honest, it's probably a lot simpler explanation like the network periodically adjusting the load.
Occam's razor.
maratsade wrote:To be honest, it's probably a lot simpler explanation like the network periodically adjusting the load.
Occam's razor.
Yes, excellent point about the load balancing! Although I don't know much about how this works, it seems that the bandwidth of the beam (at least for download) is divided among the "Out Routes". I noticed that not only has my IPGW not changed since the 11th, neither has my outroute, and I think that normally changes fairly often, like every day or two at some seemingly random time. So maybe something that does the load balancing has stopped working on my beam and people who reboot are getting crowded onto a busy out route, while out route 4 becomes increasingly empty! I was going to reboot my modem last night, but held off when I began to suspect something -- even though my evidence is flimsy at best.
I could reboot tonight and see what happens -- but if this wild hypothesis about load balancing is remotely correct and I end up on a crowded out route then I will be very sad :-(
Of course, maybe the fantastic performance I am seeing is just the result of the great crush of traffic from the pandemic being over and I am finally seeing the result of those gateway upgrades they appear to have rolled out earlier this year. Although the speed tests I ran last night in prime time were eye popping, it was Saturday evening, so things are a little different for Saturday. Sunday usually appears to have more activity in the afternoon, so maybe I can retest then to see what happens.
We could also ask that fellow with the issue for his beam, IPGW, and out route to see if there is anything even remotely lining up, after all, if he is in Utah then there can not be anything here right?
So if we ask the fellow with the issue for beam/IPGW/out route, and he is on my beam with a different out route, and I see crazy high benchmarks consistently over 50Mbps again this afternoon which is usually fairly busy on a Sunday, and if I reboot to find myself in purgatory at 7pm tonight, then maybe there is something to this load balancing idea?
If the first two look to be true, I might not have the courage to reboot though.... :-)
What do you think?
Rebooting always pops you into a lesser used IPGW. It will also have the benefit of refreshing the DNS cache, which is always good. Sometimes that thing gets corrupted. I usually end up rebooting at least once a week, because my wife likes to work downstairs, away from the modem, which doesn't always have a good wifi signal. That's usually a good formula for a busted DNS cache.
Yes, exactly -- but if that process quit working for a bit it might be a problem as on reboot people might begin to accumulate on one IPGW. There may be a similar process with Out Route?
Anyway, from his testing, I see a third city show up that is 140 miles from the approximate beam 82 center. With my testmy results, the locations were within 60 miles of beam center 2/3 of the time, and the rest were within 90 miles. Maratsade had a larger distance, but with 140 miles (and that is if he is actually beam 82, so maybe the location data is even worse than this!) then maybe test scores are not just mixed in with J1 beams, but also neighboring J2 beams.
Also, I have been running test this afternoon, and they are not above 50Mbps, but looking good in the high 30's to mid 40's, so it is beginning to appear there is nothing to this and that I just had a streak of really high tests for a bit.
I also do a regular reboot, the early patches would get squirrelly after a week, but they seem a bit better now, so I try to reboot every two weeks. I don't think I need to worry about rebooting this evening and finding myself in slow-ville.
Do you think I should ask the fellow to confirm his beam/IPGW/Out route? I don't want to waste his time with a wild goose chase.
Can always see if the guy will respond.
I had noticed a few days ago that TestMy was showing an average test score for many towns in my beam that had dropped a lot in the last few weeks, and I thought maybe something was up, since it would take a lot of bad tests to cause the average to drop so much for the towns I was watching.
I pulled the list for several towns, and copied it out to a spreadsheet and filtered it down and I am seeing averages for each town much closer to the average numbers that I thought TestMy showed a few week ago. Other areas seems to show averages similar to what they had before, so maybe it is better to not rely on the average they show, and just pull the list out to see if things are still good for the region in general using a spreadsheet. A bit of a pain as it is easy to watch a group of towns by saving a direct link to each page, and takes a couple minutes to copy to a spreadsheet to verify the numbers....
My test scores are still high, but at least the low averages in the towns with my results do not appear to be caused by many low scores showing up recently and it appears rather than being a newly cropped up issue with the beam or gateway, that something changed with the way TestMy calculated the average in these towns.
Maybe TestMy excludes scores from the Town/ISP specific average they think are too high, and something happened to exclude a bunch of the better scores for Hughes in my area? Maybe "recent" changed to include everything in history?
