What does the web response test mean?

[MrBill-2552]

I just got HughesNet service installed and found it to have slow response.  After an initial web response test result around 2 seconds, my system got two results of 0.1 seconds in the following 2 days.  What is the start and stop of this test? Does the test exclude the transport time from modem to gateway and back?

[maratsade]

The test measures the ability of your system to display a "fixed size"  web page in a reasonable amount of time. I'm not an expert, but in my opinion the test isn't very helpful, because the pages you pull up when browsing all vary quite a bit in size. The test is probably helpful if it loads up the fixed size page very slowly, but otherwise, it's probably not very helpful at all. 

If you're seeing slow response, you may want to run a test using Testmy.net.  Create a free account there so you can keep a record of your tests, and use the manual setting. Sizes should be 25MB to test downloads and 4MB to test uploads. 

The reps on this site only accept tests from Testmy.net, so you may want to restrict your testing to that site. 

Keep in mind a million things can impact your browsing speed, and many of those things are beyond our control. There are things you can do like get an ad blocker, so you will limit a bit the amount of data that the site downloads, but there are still a lot of things that slow down the works, especially during prime time.  

[MrBill-2552]

Thanks for responding. I'll check out the responsiveness next time I'm there and see if it's better now that the test is reporting 0.1 seconds as opposed to the prior 2 seconds. 

I'll also check out testmy.net when I'm there.  Thanks for that suggestion. 

It would be good if HughesNet defined what is included and not included in the test.  With a reported "web response time" of 0.1 seconds, the transport time must be excluded since it takes light about 2 seconds to traverse 88,000 miles (the distance from the modem to the satellite to the gateway and then back. Unless they have a way around that transport time - it would be interesting to know how that might be done.

[GabeU]

Actually, the latency due to distance is only around 500ms or so (half second), and that's all four legs of the trip, as the total distance is only ~90,000 miles or so, though the precise distance depends on your location and the location of your gateway (which is always in a different state than your own). The equipment itself ends up taking the total latency up to about 600-650ms on average for geo-sat internet like this. The result of 0.1 seconds on the responsiveness test is a head scratcher, as even if it included only the two legs of the return portion (from the gateway to the satellite to you) it should still be more than double that just from the latency alone. However, whether it's actually supposed to include the latency, and if so, what parts of it, I can't say, as I don't know the test's technical aspects.

To be honest, I don't even know why they have the web responsiveness test anymore, as it was primarily used for the legacy plans on the older satellites, as in the ones with a top speed between 0.5 to 1.5Mbps or so. It also possible that's why you're seeing a result like that, though I can't say for sure.

As a measurement tool, I'd rely more on the testmy.net results than anything else. Fast.com can also give a good result, but it causes the Video Data Saver to kick in, which will limit the download speed result to around 3Mbps or so, as the test mimics streaming, which is what the VDS is designed to help save data with. If the VDS is paused or turned off, Fast.com should show the actual top speed at the time. You can see what I'm referring to here.

[MrBill-2552]

Thanks for the response.  Yes, the latency due to distance is about a half second - thanks for the correction. I did the math wrong and had calculated the inverse of the latency - oops.

Perhaps my 0.1 sec web response time test was the result of my misunderstanding the test. I ran test from outside the HughesNet system (at a different location on a cable internet system). I thought the test would ping the modem on my account, but after thinking about it, the way the test is described, it is looking at the time to load a fixed page size so it needs a browser and that should have told me that I can only run that test from a browser behind the modem.  Do I need to be behind the HughesNet modem to get a valid result?  The speed test does not seem to require that. 

I did re-test the web response test when last using the HughesNet system (on Friday morning) and the response for that test reported results varied in the range of a few seconds.  Some web pages that I separately tried to open never opened.  The performance seemed appallingly slow to me.  Some phone apps would not even start.  I ran various speed tests (speedtest.net, testmy.net, HughesNet) and got results around tens of kilobytes per second.  I called tech support - the agent re-booted the modem and ran diagnostics. That seemed to correct the issue.  It will be very difficult for me to use the system as I need if frequent re-boots are needed as I am usually not at the location of the HughesNet service and I need it to  stay online and be responsive enough to operate several internet devices that I have there.

Should I expect frequent (i.e. more than monthly) re-boots?

[GabeU · Accepted solution]

Just like with a phone, a TV, a router, etc., a reboot can often solve problems and make a device run more smoothly, but no, it's not something that needs to be done on a regular basis.