New Hughes peak period performance record

[GW]

I was doing some analysis last night and arrived at a function that told me I should update Silverlight because of security issues. Since I was logged into an investment account, and my internet was basically non-functional for any practical purpose, I thought I'd go ahead with the download and come back to work another time.

Out of habit, I checked the progress of download in the browser and thought... "Really? Another 49 minutes to download 12.6 MB? Lets record this performance."

Screenshot taken at 10:34 PM and saved "Yesterday, ‎September ‎02, ‎2016, ‏‎10:35:13 PM"



I came back in about a half hour or so to see how it was doing. At 11:08 PM it was more than half done. File saved "Yesterday, ‎September ‎02, ‎2016, ‏‎11:09:14 PM". Making progress, only 12 more minutes to go.



Minor miscalculation in the projected download time. After coming back to check a couple of times, I saw the "12 minutes" part wasn't happening. I almost missed the grand finale but arrived in time to watch the last of it tick down and grabbed the final working screenshot. File saved "Yesterday, ‎September ‎02, ‎2016, ‏‎11:33:41 PM"



And after hitting the finish line. The end result, an hour to download a 12.6 MB file




How does this amazing performance rank in history? 2016 HughesNet is providing stiff competition for the early 90's 28.8 modem which would have required an extra minute of download time to get the job done. Impressive.

Yeah, I think Hughes has definitely oversold my beam.



You may wonder, am I in FAP? Nah.




I used the happy, confident, thankful, excited SENTIMENT icon but what I really mean is "Amused". Useless internet but high entertainment value nevertheless. You just can't make up stuff like this.

I sure hope Hughes isn't contracting with KaboomX to launch the next supply of rapidly depleting bandwidth.

[Gwalk900]

Just some thoughts,

Your file transfer speeds are not necessarily indicative of your Hughesnet connection speeds.

You really have no way to know what the server load and speeds were from the Microsoft download site.

Time to run a series of speed tests at testmy.net and post the results here.

Do a manual 12 MB download while connected directly to the modem.

[Knerkin Akin]

I've noticed recently that the Microsoft update servers are way overloaded. Was doing a Win7 update via the network here; 35 MB was predicted to take 35 minutes. While it was downloading, I fired up a 40 MB FTP download on another machine on the same network from our server; 2 minutes. SO: it's not always HughesNet.

[GW]

Yes, I know M$ server load is certainly a factor. It's doing better today.



Edit: testmy is currently reading me at 4.7 Mbps. It's still Saturday morning here.

[GabeU]

AS others have noted, you can't use Windows' servers as an indication of your speed.  It took me over an hour to download a single update from Windows updates that, according to my actual speed at the time, should have taken under a minute.  If I went by the Windows servers, I'd be getting slightly better than dialup, while this clearly was not actually the case. 

And again, as others have mentioned, a few tests at testmy.net, with a single device connected directly to the modem, without the router, and using the manual 12MB test size, will give you your actual speed at that time.  Speeds fluctuate throughout the day. 

[debbie.jean.bro]

I've always disliked Silverlight for that very reason and try to avoid watching anything run on it like the plague. Even when I had Comcast updates took forever and frequently just quit and I'd have to start all over. When I had Vista OS and Comcast it never did install correctly. I moved way out in the country and have Hughesnet now. I set all my updates to run during bonus bytes hours, so if something does take forever, at least I don't lose my anytime data.

[GW]

I'm well aware a flood can't flow through a soda straw. This was meant to be a humorous post.

Obviously my Hughes system is not as consistently slow as a 25 year old 28.8 dial up modem or even as bad 1xRTT cellular but at times it's so close the difference is insignificant. With testmy, I see download speeds from as low as 114 kbps up to as high as 44.4 Mbps. Gabe, I posted a 234 kbps result from last night to your current question thread.

This situation has been tested and analyzed to death, month after month and with numerous engineers involved in the analysis. Miss Amanda, over the years (no typo, that is years) has worked her butt off trying to help resolve this and another (fortunately resolved) situation for me but in the final analysis, it is what it is.

Although I don't necessarily believe it, the company finally admits there is nothing that can be done and this is the way it will be until more capacity is added.  It is high traffic causing my slowdowns and presumably causing the same poor performance for all us poor sots stuck with Beam 55/RAP for service. At least that's what they want me to believe.

It's not just download speeds either. I see primetime upload speeds that are even more ridiculous. I see testmy upload results as low as 4 kbps. Below that, the test simply fails.



There was one recent period when I had a hard time breaking 20 kbps upload using the Hughes Performance Test. 14 out of 20  tests over 11⁄2 weeks came in at 21 kbps or lower.



Regardless of major work being done on the Rapid City gateway this summer and the use of all the load balancing tricks they could employ, nothing will change for me until they launch the new satellite and free up some space.

I know I'm editorializing but I find it hard to believe a company allows such gross mismanagement of assets. Sadly, it seems as though totally overselling the asset and greatly impairing the service of those who have no viable option is what puts new birds in the air.

[GW]

Stellar performance again tonight.

[GabeU]

GW, what is your results URL so your tests can all be looked at?  If you don't have one, you should create an account so all of your tests will be saved. 

[GW]

I usually don't bother logging into testmy since their recordkeeping seems really screwed up to me.

For permanent records, I use the Hughes test.

I did log into my testmy account tonight after my earlier post. I wanted to touch off a series of automated tests every 10 minutes to see how consistently poor my internet is this night. That just finished.

[GabeU]

With regard to their record keeping, do you mean when it includes tests that you didn't do?  If that's what you are referring to, on your results page there is a thing to change on the bottom left of the results field.  Something regarding only displaying the tests from YOUR connection ID.  Change that, and you will only see your tests from now on.

As for what the reps will need, right now only your Hughesnet performed tests will be accepted, but at least you've done them there, so they will be able to look at the results.  The reason for this is that in order for the reps to accept the testmy.net tests as valid, they need to be done with the manual 12MB size for download tests and the manual 2MB size for upload tests.  But, you stated that you've been doing tests at Hughesnet, so they'll have a record of those. 

[GW]

Yes, that's exactly what I mean but it's supposedly set to display results for my username only. That feature doesn't always work as designed.

The problem with my internet isn't a lack of supporting data. Everybody knows it and sees it. The problem is there is nothing that can be done about it.

[GabeU]

How do you know there is nothing that can be done about it?  Have you created a problem post to see if they can do something?  I looked on your profile and didn't see any.  The only relatable thread was a question post regarding upload speed.   

[GW]

How do I know? I wrote the post above that you may have missed. Please read the one that mentions your name.

[GW]

Maybe there is something. Maybe my modem is melting down. Besides the creepy-crawly performance that seems to grow worse by the day, the modem shuts down if I even touch it the wrong way. There was an interesting thing that happened a few weeks ago too. The performance gradually slowed over a few hours until I couldn't even load a page or access email. A reboot of the terminal was required to make it functional again.

It is really strange to consider Monday morning performance like this can be attributable to high traffic loads. This is two different computers using two different ethernet cables with both wired direct to the modem when tested. Automated test routines set to run every five minutes, ten times. I set the second test session to run 12 mb down and 2 mb up.


.

[GabeU]

You started this thread two days ago.  So, in the two days since you started this thread, without any input from the Official Reps, who aren't on here on the weekends, nor on holidays, nor the engineers, you've automatically come to the conclusion that "there is nothing that can be done about it." 

Or, again, and perhaps I just missed it, where is a prior thread in which you brought up a speed issue, the Reps and engineers addressed the issue, and the ultimate conclusion was that "there is nothing that can be done about it?"  I've yet to find anything to justify that conclusion, but I'd be happy to read it if you could post the thread.   

[GW]

Sorry for the confusion. The conclusion was made by Hughes at the end of July. You won't find any discussion of this in the forum.There is none.

I didn't intend to open up a discussion for advice or information from members or representatives. My original intent was merely to post up some humor about the often absurd nature of HughesNet service.

The follow-up information I posted may be of some interest to others who are stuck with Beam 55 and the Rapid City gateway.

[GabeU]

Then hopefully the Reps, by reading this thread, will be able to address your problem.  Remember, they have a lot more "power", if you will, when dealing with the engineers and such than the phone reps, and there may very well be some tweaking that can be done.  I also think that the reps on here work much more diligently than the phone reps, including "level four support", or whatever it's called, in seeing the solution through.  All hope is not lost.  There is a particular customer who was having terrible speed problem by the handle of pbparker.  After a few weeks of tweaking by the engineers, his speeds are now very respectable.  It took a little trial and error from what it sounds like, but they eventually got it.  You can go back through and find his thread within the last month of posts. There may be some type of load levelling or something else they can do, even if your beam is packed. 

[GabeU]

The second set of test will most certainly help more than the first.  But, with that said, they may ask you to run a few tests (not that many) at various times of the day to get an idea of patterns and such. 

With that said, if you do the automated tests again, at the various times of the day (and I didn't realize you could set the speed, time space, and number of tests, so that's a great find for me), I would use the same size that you did in the second tests, set them to run five minutes apart,  like you did, but set it to only run five tests.  What you did is great, but it's really overkill, and you don't want to waste precious data.  No one wants to use more data than they have to.  A block of five tests, each set five minutes apart, and a block at various times of the day, preferably one in the morning, afternoon and evening, if you can, is really what would be the most helpful. 

Again, hopefully there is something they can do.  There may very well be.   

[GW]

I appreciate the advice Gabe but running five 12 MB/2 MB tests on a 5 minute interval during primetime doesn't work too well. It can take 10-20 minutes to complete a single test. It sets up continual speed test mode. A test project that should complete in mere seconds split five ways uses up more than an hour of computer and internet time. For this particular example, I began the five test/five minute interval session at 8:13 PM. It completed at 9:31 PM. This is why I tend to use the defaults which choose little doinker file sizes more appropriate to the testmy perceived level of performance.




The Hughes test is quicker but when service is impaired, it never even tries to do the 4 MB download test.



Spending data doesn't bother me though. I have tokens from 2010 I've never used.

[BirdDog]

GW, You definitely have a problem, speed should not be that low even in peak prime time. Looks like what I had several years ago when had a problem. Highest I was seeing depending on time of day was maybe around 2 Mbps.

Which modem and plan do you have? You are plugged directly into the modem?