Pelican Parts Forums

Pelican Parts Forums (http://forums.pelicanparts.com/)
-   Off Topic Discussions (http://forums.pelicanparts.com/off-topic-discussions/)
-   -   Testing Internet Bandwidth/Speed Sites (http://forums.pelicanparts.com/off-topic-discussions/607487-testing-internet-bandwidth-speed-sites.html)

Rusty Heap 05-08-2011 03:05 PM

Testing Internet Bandwidth/Speed Sites
 
Anyone have some independant websites to ping to test bandwidth, ping response, and download speeds?

Normally I've used Speedtest.net - The Global Broadband Speed Test

I'm wanting to get some other sites to check accuracy and comparison.

I'm on a Hughesnet Sat Dish for web service, and honestly don't believe thier own internal "speed check" numbers. I just upgraded to their "fastest" service which is supposed to average 1.5-2 meg bits a second. I'm calling bull on thier speed test that says that's what I'm getting, as my home internet is still dead dog slow.

On speedtest.net, I'm consistantly in the 500-700 kbps range.

Anyone have other suggested independant web sites to do some comparison testing of my bandwidth?

Paul_Heery 05-08-2011 03:15 PM

Try this site. Use the "Huge Download test". Downloading an actual file a few times will give you a good benchmark
Download Speed Test

masraum 05-08-2011 04:19 PM

Do a google search for "speed test" and try 3 or 4 different sites and try each one 2 or 3 times.

I used to work in the Commercial Satellite industry (voice and data to oil rigs, ships, trucks that drive around to out of the way places, etc...). Satellite bandwidth is very expensive. Are you using Satellite for download and a phone line for upload, or have they managed to make the home Internet service the sat link bidirectional these days? I hope/assume they have. Years ago, I knew someone who worked for Hughes DirectPC, it used a phone for upload. There are several problems with Satellite.

1 The outbound signal is one huge signal for everyone (or at least, everyone in a certain region). Everyone in that region downloads the entire signal (including the data for everyone else) Once you receive the signal, your data is tagged with the address of your sat box and so you only see your data.

This works really well for TV (which is what this stuff is really setup for) because they send a ton of data, and you pick which data you want to see. That's why you can receive HD programming (I think the requirement for HD is 6-10Mb).
http://forums.pelicanparts.com/uploa...1304900300.jpg

2 Also for data (rather than TV), they have several tricks that they do behind the scenes to compress the data and proxy the data to make the two sides believe that they aren't talking over a link that has a latency of ~600ms (anywhere from 500ms up to 800ms depending). Sometimes this proxy action can make a huge difference. It usually only works for TCP connections, but not UDP connections. Therefore, if you're doing something that involves a UDP connection, you probably aren't getting the benefit of their tricks. The computer sees that the delay between your computer and the Internet is over half a second round trip, and slows stuff down to compensate for what it thinks is a slow link. Most PC communications is set up for a link with low latency of pretty much any speed. A high speed link with high latency often confuses a PC (which is why they employ the trickery). There's not anything that you can do about the long ping times. It takes radio waves a certain amount of time to travel that distance. There's nothing that you can do about the laws of physics.

3 As I mentioned above, the outbound is shared between everyone. If you're using at peak time, you'll see much lower speeds than at an off peak time. I suspect that you may have a guarateed minimum of X (maybe 512k) with a possible peak of 1.5Mb. When you test at a peak time, you'll see that minimum.

4 The inbound signal is much, much slower. Instead of everyone sharing one massive signal like the outbound (which is possible because they have 1 massive dish with a hugely powerful transmitter serving up the outbound), everyone gets a tiny inbound signal (everyone has tiny little dishes with low powered radios). Most Internet server relies on relatively small queries from your machine to request larger, peaky amounts of data from the Internet. The system is setup for that sort of thing. Also, lots of internet usage is bursty. You spend a few minutes reading a page during which time you aren't downloading any data at all. During this time, your neighbors are using your data. Then, you click a link and you ask for a web page to be downloaded. Hopefully during that time, your neighbor is now reading the page he downloaded.

http://forums.pelicanparts.com/uploa...1304900362.jpg

You may want to check what your speed is supposed to be (minimum guaranteed is important). And check in several locations and using several different mechanisms. If a test uses a UDP download, it'll appear much slower than a TCP connection (like FTP).

Rob Channell 05-08-2011 07:09 PM

http://speedtest.knology.net/

The site for a local provider....

Radioactive 05-08-2011 07:23 PM

Speakeasy - Speed Test

Rusty Heap 05-08-2011 07:37 PM

Thanks Steve for the detail explaination.

yes hughes is both up/down to the Bird, no phone lines.

I find it's not just peak times, but across the board 24/7

problem is I'm out on 5 acres in rural woodlands, and no other option for internet.

I'm not a power user trying to stream movies or webtv. my ping latency averages about 800-900 milli-seconds.


thanks all for the web links, will try them all out for an averaged speed test(s)

masraum 05-09-2011 04:59 AM

Quote:

Originally Posted by Rusty Heap (Post 6010668)
Thanks Steve for the detail explaination.

yes hughes is both up/down to the Bird, no phone lines.

I find it's not just peak times, but across the board 24/7

problem is I'm out on 5 acres in rural woodlands, and no other option for internet.

I'm not a power user trying to stream movies or webtv. my ping latency averages about 800-900 milli-seconds.


thanks all for the web links, will try them all out for an averaged speed test(s)

One other thing that I forgot to put down. Your dish has to be pointed right at the satellite. If the dish is not pointed precisely, that will affect your throughput. When the satellite transmits down to the ground, it does so like a flood light trying to hit a large area, but when your dish transmits up, it is more like a spot light trying to keep the "beam" as tight as possible. If that beam isn't pointed with precision, it will reduce your throughput. Also, where you are in the satellites coverage area will affect things. If you are near the edge of their coverage, you'll not get as good a signal and your throughput will suffer.

800-900 ms is definitely at what I'd consider the high end, but the satellites are over the equator. Since you're in the north, that means that the transmission has to go a bit farther and so the time would be longer. How much longer, I'm not sure. I know that we had customers with vans that did a lot of work in Canada, but I can't remember what their ping times were.

dhoward 05-09-2011 09:51 AM

Most ISP's throughput numbers are to THEIR local office. Anything past that they won't be responsible for. I learned that after a month's worth of fighting with Charter. NEVER get advertised speed.
"Oh, 32MBPS is just to OUR site..."

Mark Henry 05-09-2011 10:46 AM

I'm averaging a little better than 5 meg bits a second download on rural wireless service. The antenna on my roof looks like a small pizza box sitting on it's end.

$45/month, unlimited, taxes in.

Super_Dave_D 05-09-2011 03:50 PM

http://speedtest.comcast.net/

Mine is a little slow:)
32.5 down
7.1 up


All times are GMT -8. The time now is 11:36 AM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2025, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.0
Copyright 2025 Pelican Parts, LLC - Posts may be archived for display on the Pelican Parts Website


DTO Garage Plus vBulletin Plugins by Drive Thru Online, Inc.