Understand Your Internet Connection and Test its Speed
Since the release of DSL, Internet Service Providers (ISP) have been marketing the Internet Bandwidth Speeds of their service to steal customers away from their competitors. The first battle was Dial-Up vs. DSL. Followed by DSL vs. High-Speed Cable and of course today it’s High-Speed Cable vs. Verizon FIOS.
Honestly, the battle for “Speed” supremacy will probably never end NOR IS THAT WHAT I WANT IT TO DO. :) After all, when it comes right down to it, as the growth of the ONLINE services continues (HULU, YouTube, Streaming Netflix, Amazon Unbox, PORN…) so do the demands of Faster Internet Connections.
So, as a Groovy Reader, you are in for a treat because today I’m going to talk about:
- How to convert your ISP Marketing Lingo into English
- Where you can Test your Internet Speed
- Sorry, that’s it
So, let’s jump right into it!
How to convert Marketing Internet Speed/Bandwidth Lingo into Real English
Be an informed consumer. It’s even more important to be a GROOVY informed consumer! So let’s take a look at a Verizon FIOS Internet Advertisement and dissect it.
Note: Mbps above is supposed to be written as mbps. Verizon marketing guys should fix that.
WOW! 10 mbps Download and 2 mbps Upload! That’s amazing!! Right? At those speeds, I should be able to download an entire DVD in…
- 1 DVD = 4 Gigabytes
- 4 Gigabytes is 4000 Megabytes
- So at 10 Mbps the math would be 4000/10 = 400 seconds! WOW!! An entire DVD in 6.6 Minutes!
Actually no. You see, here’s the problem. Mbps does not = MegaBytes per second. Mbps = MegaBits per second. Let’s quickly get familiar with the Math of the Verizon FIOS Ad above using 10 mbps:
Speed | KiloBit | KiloByte | MegaBit | MegaByte | GigaBit | GigaByte |
Abbreviation | kbps | KBps | mbps | MBps | gbps | GBps |
10 mbps conversion | 10000 | 1250 | 10 | 1.25 | 0.01 | 0.00125 |
Reference: Numion.com has a GREAT Data Unit converter I highly recommend!
So the actual math for a DVD download at 10 mbps would be:
- 1 DVD = 4 Gigabytes/GB
- 4 Gigabyte = 4000 Megabytes/MB
- 4000 Megabytes = 32000 Megabits
- 32000 Megabits/10 mbps = 3200 seconds or 53 Minutes
Here are a few more examples of downloading at 10 mbps/Verizon FIOS Ad Above:
Type | Web Page | Kid’s Picture | iTune Song | CD | DVD |
Size | 100 KB | 1 MB | 3.5 MB | 700 MB | 4 GB |
Download Time | Instant | .8 Seconds | 2.8 Seconds | 9 Minutes | 53 Minutes |
Note: These are BEST CASE scenarios. Normally if you get 50% your speed, you’re doing good!
Clear as mud? Just remember:
8 Megabit (mbps) = 1 Megabyte (MBps)
Okay Geez I understand Megabits already! How do I check my connection speed?!
How-To Test your Internet Connection Bandwidth Speeds
There are a couple of places that I frequent for this purpose, and both are dead simple:
Usually, I use Speakeasy.net. It doesn’t have all the eye candy of Speedtest.net, but I like simple and clean.
1) Just Open your Browser and Type https://speakeasy.net/speedtest/ in the address bar.
2) Once the page loads Click the closest Server Location to begin.
The SpeakEasy site will test out your Download and Upload speed. When the test finishes, your Upload, and Download speeds are displayed using kbps or Kilobits per second.
In my example, my test came in at 14,740kbps Download and 11435 kbps Upload
- 14,740 kbps = 14.7 mbps which = 1.84 MBps or Megabytes Per Second
- 11,435 kbps = 11.4 mbps which = 1.42 MBps or Megabytes Per Second
Not bad EXCEPT, I’m currently paying Verizon for the 20 mbps Download and 20 mbps Upload FIOS plan. It looks like Verizon Owes me some bandwidth! :)
Personally, I’m not going to call Verizon and demand my money back. Why? Well, the problem (that we’re ALL going to start facing) is that the bottleneck with Internet access is no longer you and me. The bottleneck is slowly shifting to Websites and Internet Services. Websites like Youtube, Netflix or even Speakeasy only have so much bandwidth, so all their customers have to share what they have. Just think what happens to Speakeasy if 100 FIOS customers all hit the server at the same time to test their internet connection.
Now granted, HUGE companies like Google and Microsoft have super groovy Internet Pipes, but smaller sites don’t. And when people like me connect and try to consume a full 20 mbps, the Website/Service might not have enough Bandwidth to give me all 20 mbps. So be patient. As consumers get faster Internet pipes, so do Websites and other Online Service providers.
This article was fun to write. I would LOVE to hear about your Speed Test results! Do YOUR results match what your ISP sold you?
In your comments below, please include the following so we can compare and get some perspective:
- City and Service Provider
- mbps plan (Up/Down)
- Speed Test Results
One thing to note, on Windows machines, I recommend running the tests using Internet Explorer. Running the tests with Firefox I consistently get slower upload speeds (4000-5000 kbps slower upload). Some say it’s because the FLASH plugin for IE is faster than Firefox but I have no idea. Check on your own and post your comments below.
39 Comments
Leave a Reply
Leave a Reply
Bill
March 1, 2009 at 10:09 am
Just found your site from DIGG. Great article! I actually had no idea there was a megabit vs. megabyte.
Anyway – here are my stats:
Seattle WA with Comcast
No idea what I pay for….
12832 mbps up / 5393 mpbs down
shockersh
March 1, 2009 at 12:53 pm
SanFran : Starbucks on WIFI
1833 mbps Down / 1298 Up
Not bad for free WIFI I guess ;)
shockersh
March 13, 2009 at 12:25 pm
Update – Ran this test from work today:
24922 kbps Download / 35029 kbps Upload
Upload faster than Download… I can only assume this because we have huge pipes for our website….
Alexmvp
March 2, 2009 at 11:58 am
Tested from work (smokin pipes!) and I didn’t have the same issues you reported in regards to IE Vs. Firefox:
Firefox – 31877 kbps Download and 36327 kbps Upload
IE7 – 31464 kbps Download and 36902 kbps Upload
Animis
March 2, 2009 at 3:09 pm
Tested using the Seattle server, My ISP is comcast, paying for 12Mb/4Mb
13559 down
5899 up
A little better than what I’m paying for!
Blake
March 2, 2009 at 3:54 pm
Bottlenecks don’t only pop up at the webserver level, but also with the ISPs when many people are online at once.
Also, you used a ratio of 1000 when going from KB to MB, MB to GB, etc (like 4 GB = 4000 MB in your DVD example), but this is slightly off. A kilobyte actually has 2^10=1024 bytes, which is approximately 1000. A megabyte has 2^10=1024 kilobytes, or 2^20=1,048,576 bytes, etc (http://computer.howstuffworks.com/bytes3.htm).
MrGroove
March 2, 2009 at 7:48 pm
@Blake – Good point. My understanding is that is the Binary definition (1024 bytes per kilobyte) as used for data storage but not for expressing bandwidth or throughput. Something that probably could have been explained a bit deeper by grooveDexter.
shockersh
March 13, 2009 at 12:26 pm
@MrGroove, yeah – everytime I format a PC drive and create new LUNS or Drives, I always use 1024 as my multiplier. I didn’t think that applied to throughput. Thanks for FYI.
MrGroove
March 2, 2009 at 7:49 pm
@shockersh – ick… that’s a pretty slow connection.
@alexmvp – wow, those are sick speeds! You must work for an ISP :)
Jason
March 12, 2009 at 6:20 pm
THANK YOU for explaining this. I work for a small ISP and we have to constantly talk to people who think they aren’t getting what they are paying for.
MrGroove
March 12, 2009 at 6:47 pm
@Jason, Hi Jason Welcome to the site and Thank you for the feedback! It’s great to hear from an ISP on this topic. Feel free to link to the article from your FAQ on your ISP Pages. :) :) :)
Hope to see you around!
TechSlave
March 12, 2009 at 6:54 pm
A few more points, more in-depth, on this issue:
‘Speedboost’ technologies, rolled out by Comcast as well as some others like Time Warner cable (Powerboost).
These basically prioritize the first 10MB of a download or upload which can cause speed testing results to be much higher than actual averaged speeds. Though its not listed in ‘megabytes’ in the customer facing documents, it simply says the first 10-15 seconds of file download(or upload) is boosted. This of course relies on bandwidth availability.
Mind you, some companies also use ‘on network’ speed tests rather than ‘off network’ so that their internal pipeline rather than external services are the bottleneck. This provides better information in some cases, unless your area is very geographically distant from the on-network server doing the speed test.
Another thing to remember on speed test results is peak usage, especially on shared pipe (cable) systems as compared to fiber or DSL. I deal with speed-related complaints for a major cable ISP and between 6pm and midnight, kiss full bandwidth availability goodbye as everyone logs into Xbox live to stream Netflix, hits those viral videos they couldn’t watch at work, goes to MySpace and hits twenty pages in a row loading audio streams…etc…etc. Not to mention P2P applications.
With DSL or Verizon, your speed test results will be less impacted by on-network clogging (bandwidth shared on the ISP networks) than it will be by off-network issues (server, webserver, and hosting bandwidth capacity).
A final note: Speedtest.net has more providers for their speed-tests, and it definitely makes a difference. From work at the ISP, I can find speeds anywhere from 5megabits per second to 35 megabits per second depending on whose server I am testing to and how the backbone routing, peering agreements, and intermediary providers are set up.
MrGroove
March 12, 2009 at 7:00 pm
@TechSlave, Great Feedback TechSlave. Thanks for the great information drilling further into the details on this. I’ve never worked for an ISP so the article was really missing these types of insights! Welcome to the site and I hope to see you around more!
TechSlave
March 12, 2009 at 7:48 pm
@MrGroove:
No problem, it can be pretty confusing when dealing with speed issues. You hit an important point – the megabyte/megabit transition can be really confusing for people too. From the ISP side, there is definitely a lot of things that have to be checked.
Since I’m on the phone with a customer about this kind of issue right now, I’ll throw out a rather wordy run-down on speed problems, at least from the most-common demoninator viewpoint:
Some of the crazier issues are problems with the TCP/IP stack itself which require either a TCP optimizer (also known as an invitation for your friend BSoD to visit and beat up your Windows PC for awhile), or a reset of the TCP/IP stack.
Parents often call with speed issues that turn out to be uTorrent or BitTorrent run by another user on the computer, eating all the bandwidth because nobody throttled the torrent program from using all available bandwidth. Other program-related problems can be antivirus programs, or dueling software firewall programs.
Generally, when you call an ISP to run speed-tests when you aren’t getting a speed near to what you are purchasing, make sure you’re prepared.
1. Go direct. Modem->PC. Do not be wireless. If you can, use ethernet rather than USB connection. Routers and wireless connections can all impact your speed, especially with unsecure wireless connections where someone is leeching – or your 2.4ghz cordless phone, microwave, or another radio frequency contamination is breaking up your signal.
Additionally, make sure to restart the modem, then reboot the PC. Some modems have batteries – make sure the modem fully powers down and re-connects.
2. Turn off your software firewall programs. Most firewall programs such as Norton or McAffee can be turned off for a specific period, in case you forget to reactivate them after the speed testing.
3. Turn off any P2P/torrent programs. They’re just bandwidth hogs if you don’t know how to manage them.
4. Make sure your antivirus program is not currently running a scan, which can eat RAM/CPU processing cycles so that your pages load slowly.
5. Clear cache on your web browser. If you can, test with two – there are some subtle types of corruption that Internet Explorer can experience, especially related to add-ons, that may be eating your bandwidth.
6. If you don’t have firewall, Antivirus, and anti-spyware – download and run antivirus and anti-spyware. Many major ISPs provide free antivirus and internet security suites with your service. Clean computers are happy computers!
There are some other thoughts, but really, this covers 75% of the speed issues we get calls on. Another 20% is related to radio frequency problems – generally issues with cabling, ethernet cords, modem problems. The final 5% is network congestion which is difficult to resolve.
MrGroove
March 12, 2009 at 8:46 pm
@TechSlave, Again, great information. A quick point for other readers – When you say “MODEM” your talking about the DSL Modem or Cable Modem NOT old school 38.8 BPS Modems. For old-schoolers like me, Dial-Up Modem was the first thing I though about when I read your article.
I really like your point about making sure you have control of your Internet connection BEFORE you test your speeds or call your ISP to inquire about your Data Speeds. I personally ran into this when I was invited over to my Sisters House only to discover it was actually a “PC Support” Dinner…. (I’m sure all PC guys have had one of these). Anyway, turns out my nephew had installed BitTorrent on his PC and it was sucking up all their Upload Bandwidth due to the “Warez” he had downloaded and was now unknowingly sharing.
So very good point. Thanks for the great list! I’m sure the other readers will really appreciate it!
TechSlave
March 12, 2009 at 8:53 pm
@MrGroove:
Tech support dinner is definitely familiar. Had to set up my dad’s DSL and wireless and then my mom’s PC needed cleaning. Physical cleaning. That was a nightmare. Cat and dog hair in the power supply fans. And one stick of RAM was missing. I found it at the bottom of the case, and I don’t even want to know how it ‘fell out’ of the clips. Aaah, amateurs.
But yeah, I mean broadband modems. Dialup modems are their own special nightmare, or services like Wild Blue or other satellite-based stuff. My old dialup 2400baud is someplace in my mom’s basement.
Glad to put some info out there on this – one of these days I will have to write up a whole intermediate level document on this issue with more detailed information since most users wouldn’t know what switches to use with netstat, how to reset TCP/IP on Windows XP or Vista, how to reset Winsock, etc. But I hope it helps your readers.
shockersh
March 13, 2009 at 12:24 pm
@TechSlave – lol Tech Support Dinner! I never knew there was a name for it! I love it! btw – l guess that kinda goes along with the name TechSlave ;) Great write-up. You should join the crew and write up a few groovy articles TS.
Later! It’s Friday and I’m outta here. Gonna cruise the Forum first of course. ;)
TechSlave
March 13, 2009 at 7:41 pm
@shockersh, it beats going out on a ‘Tech Support date’ by a long run. I’ll think about writing things, but it would impede the free flow of my philosophically motivated ‘slacking’ and that would be a capital crime in the world of Dobbsian geekdom. J/K. But if I run across something I feel compelled to write on, maybe I’ll toss some info up.
MrGroove
March 13, 2009 at 10:29 pm
@TechSlave, LOL – well I can happily say I’ve never had a ‘Tech support Date’ of course but MrsGroove probably has something to do with that.
Anyway, welcome to the Community TechSlave, looks like you already have a few fans. ;)
MrGroove
March 13, 2009 at 10:31 pm
@TechSlave, Cat and Dog hair…. brutal! And yes, I appreciate the informative comments. Glad to see I’m not the only old-schooler out here with a 2400 Baud modem in the basement!!! Reminds me of my old BBS…
TrajkLogik
March 18, 2009 at 6:48 am
Nice read. We have a T1 here in Miami (unfortunately Fios isn’t available in our area) and we’re getting good speeds (using speakeasy) to Atlanta and Dallas (D-1500kbps U-1500kbps), but the upload speed to Seattle is only half of what it should be (800kbps), every time I test it. Does anyone know why this might happen? Thanks.
MrGroove
March 18, 2009 at 9:30 am
@TrajkLogik, Seattle is a LONG ways away and therefore you will encounter latency. Latency is the time it takes the data to get from point A to point B. Unfortunately there’s no ways to get around the Laws of Physics so the data can only travel as fast as the speed of light aka – Latency.
MrGroove
March 18, 2009 at 9:31 am
@MrGroove, Some Websites / Services get around this by putting Caching Services in place around the globe. Limelight and Akamai are examples of this.
TrajkLogik
March 18, 2009 at 9:46 am
Thanks for the reply Mr. Groove. But, why would latency only affect the upload speed? The download speed is normal.
MrGroove
March 18, 2009 at 3:17 pm
@TrajkLogik, oops… leave it to me to only read half the question…
My suggestion would be to test it during different times of day to rule out server load however still, it’s a long way to go and i’m guessing mileage will vary. :) Test different servers from around the US and pull an average. For your T1, at least your not sharing the link BUT, your carrier will probably only give you an SLA on your speed based on a complicated algorithm to their POP and not to a website 6000 miles away :) Good news is, perhaps your a bit more educated now due to this groovy article and you know how to ask the question to get accurate data. :) ???
speeddemon
April 1, 2009 at 8:11 am
Well I think I have you all beat!
Download Speed: 79184 kbps (9898 KB/sec transfer rate)
Upload Speed: 10341 kbps (1292.6 KB/sec transfer rate)
I’m in Chicago and I work for an ISP :)
Mauricio Feijo
April 10, 2009 at 7:28 am
What is funny is the date when you posted that! :)
Now seriously, that is smoking fast!
MrGroove
February 19, 2010 at 9:39 am
Aye – that is smoking fast but… my home upload speed is still rockin over your speed ehhehe ;)
daniel
June 24, 2012 at 8:42 am
my top download speed 5.7MB/s
top upload 3.7MB/s
and i play counter strike online with 30MS ping
Steve Krause
June 24, 2012 at 8:48 am
Very nice. Is that a cable modem or Fios? Or?
BE Durocher
February 19, 2010 at 7:49 am
Download Speed 5763 kbps, Upload kbps 2290. This is wireless and is a dog. Way too slow. I’m in Bradenton, FL on vacation and hope it will work faster at home. Annoying. Thank you. I don’t know the provider. The speed fluctuates and said 54 now 48 Mbps, Thank you.
MrGroove
February 19, 2010 at 9:39 am
5763 isn’t to bad honestly especially from a wireless link on vacation. I’m guessing you’re at a hotel or resort? Those can really be DOG SLOW which is lame being that normally you need to pay $10 – $15 for the internet connection…..
adasda
August 17, 2010 at 10:49 am
mine says 10mb but ive never downloaded anything faster than 1mbps :/
MrGroove
August 17, 2010 at 1:05 pm
Take a look at the article tho. I think what you mean is it shows 10,000 kbps correct and you’re talking about downloading from sites getting getting 1MBps?
Sarah @ Atlanta Wireless 4G
January 17, 2011 at 8:15 am
Hi, just wandered by. I have a Atlanta 4g site. Amazing the amount of information on the web. Wasn’t what I was looking for, but good site. Take care.
Marc
June 11, 2011 at 3:12 pm
Interesting results from speakeasy.net. Ran my first test series w/ Norton firewall down and got very poor results… ran second test series w/ firewall back up and results markedly improved ( note – both tests reported as average of 10 individual runs). Thought the reverse would be the case…
Tested Seattle server via Comcast ISP running 20Mbps speed tier;
Test 1 (firewall down) = Download Speed: 13552 kbps (1694 KB/sec transfer rate)
Upload Speed: 4481 kbps (560.1 KB/sec transfer rate)
Test 2 (firewall up) = Download Speed: 22858 kbps (2857.3 KB/sec transfer rate)
Upload Speed: 4561 kbps (570.1 KB/sec transfer rate)
More notes – to optimize results, all tests run over Windows 7 OS on Intel i3 CPU/2.13GHz processor/4 Gigs RAM w/ direct connection to DOCSIS 3.0 modem (disabled my wireless connection for this), reset my TCP/IP stack from command line (netsh int ip reset c:
esetlog.txt), flushed my DNS cache, reset IE settings.
jessica@cable internet
June 24, 2011 at 12:12 am
Some speed test sites lead users to believe that testing off a server close to them is necessary. This isn’t always true.Their are speed test servers are configured to maintain quality of service for thousands of miles. Most internet consumers don’t think about the internet service providers responsibility in peering beyond their own network. Many internet providers send users to their own internally hosted connection test.
Sheila
April 30, 2012 at 8:01 am
Graham, NC – Time Warner
pay for 30 mbps down and 5mbps up
get 30mpbs down and 4.28mbps up on Speakeasy.net
I’m happy.
mike
February 10, 2013 at 10:10 pm
Wise VA comcast pay for 14download and 5upload. I get 15.41Mbps download and 4.75Mbps upload