Thank you for being a valued part of the CNET community. As of December 1, 2020, the forums are in read-only format. In early 2021, CNET Forums will no longer be available. We are grateful for the participation and advice you have provided to one another over the years.

Thanks,

CNET Support

General discussion

What speed am I getting?

Dec 31, 2007 11:09PM PST

I just got comcast hispeed at 6Mbps. I've been testing with diff sites and getting widely different results. I've new to this. Please help me understand these results.

Is there a better site to use? The tests were done under identical conditions with nothing running (no firewalls, no other programs, and results are typical of repeat tests at each site).

Speakeasy
Download Speed: 10932 kbps (1366.5 KB/sec transfer rate)
Upload Speed: 1602 kbps (200.3 KB/sec transfer rate)
Used Washington DC node since I'm near there.

CNET Broadband width test on their reviews forum-
877.4 Kbps - You 877.4 kbps

Speedtest.net at http://www.speedtest.net/results.php
1.15Mb/s down .19Mb/s up 24ms latency(what's this?) <50miles
http://www.speedtest.net/result/217324682.png

I've been having a problem with my TV cable where some channels get white lines thru them (like VCR tracking lines). If my broadband cable is also poor, it will help me in asking comcast to check the lines. (Before getting the digital convertors a week ago, the TV programs would freeze for several seconds frequently instead of the white lines. When happening it's several unrelated channels, but not all channels.)

I want to add digital phone, & so I'd like the lines in good shape before I give up my current decent (but expensive) phone service.

Any help is appreciated!! As I said, I'm very new to this (I was a programmer though in a past life, so computers in general I'm okay with.)

Also, as long as I'm asking - I've added netgear 824 router & it seems fine. Any opinion between that and Belkin's 8233-4 N router? The Pre-N got rave reviews & lifetime warrantee (vs. NG's 90 days). The Belkin N1 doesn't sound great. Very little out there on the N itself.

Thanks again!!

Discussion is locked

- Collapse -
The problem here is
Dec 31, 2007 11:50PM PST

The internet is used by more than just you and your machine. A lot of sites are used by others or you get a clog now and then.

So keep measuring if you must but the answer is you should never get consistent results. This can drive some insane and they'll post across dozens of forums asking how to fix that.

Bob

- Collapse -
Welcome to the world of . . .
Dec 31, 2007 11:54PM PST

Averages. I have found the Speakeasy site to be the closest to the advertised speeds. Maybe they are in league with the ISPs. My Comcast/Time Warner/Roadrunner (pick one) advertises 6 Meg down. I average around 6.8. And I get different speeds at different speed test sites.

You have to keep in mind the size of the pipe you are accessing the sites with. Not yours, theirs. And net congestion, and the ATM switch, and the time of day, and the list goes on.

BTW. If you are paying for 6 Meg and Speakeasy says you're getting 10 meg, smile.

- Collapse -
Thanks - inconsistency still confusing
Jan 1, 2008 2:43AM PST

Thanks for the quick replies!

What's confusing me is they are repectively saying: 10M, .8M (<1M), and 1.5M. That's a huge range, so something's not making sense.

They were taken within minutes of each other. ...And are consistent to the site. Speakeasy's other one was 8M, Cnet's was .7 (and .8 a few times), and speedtest's was 1.1M.

I do know that it's a shared line, and results based on other usage will vary. However, in this case the variety isn't random; it's based on the site used to test with.

Anyone know which one matches actual width available? Any of test sites considered the "go to" one?

(Interesting comment on speakeasy matching the isp's closely - hopefully means they are accurate because that 10M did put a smile on my face : ), until I saw the others and got confused.) So far my web access is working fine, but I'd like to be more comfortable before I talk with the cable company about possible TV problems, and before I add digital.

Any ideas on my other original questions?

Thanks!

P.S. I've forgotten to include the results from this site:
Speedguide.net
SG Speed Test Results
8301 kbps down (~8.3 Mbps, 1013 KB/s) ?

- Collapse -
"Is there a better site to use?"
Jan 1, 2008 2:46AM PST

No. You tested your line at over what you were given so you are done. It is entirely PROPER to see the numbers bounce around.

Bob

- Collapse -
p.s Actually I think speakeasy is saying 1.3M
Jan 1, 2008 2:50AM PST

"Download Speed: 10932 kbps (1366.5 KB/sec transfer rate)"

That's 1366Kb is 1.3M. The 10932 with little 'k' I assume is kiloBITs.

I asume I'll have access to less than the advertised 6M, but 1.3 and 1.2 and .8M seems a lot less. Maybe indicating a line problem to go with the TV one. Any thought?

- Collapse -
Clarification of typo
Jan 1, 2008 2:57AM PST

I assumed i WOULDN'T have access to 6M, but 1.3 and .8 seem low for testing results.

- Collapse -
That's beating the speed by quite a lot.
Jan 1, 2008 3:15AM PST

The Comcast line I see is 6 Million BITS per second.

You measured over 10 million bits per second so why are you not happy?

Bob

- Collapse -
I'm not getting 10M, maybe 1M & here's why it matters
Jan 1, 2008 5:21AM PST

It's not a question of happiness. It appears that each site is very different than the next. That makes the results meaningless until there's an explanation of if ANY of them are measuring accurately (even somewhat). (In science if you're numbers are all over the place but consistent to the technique, chances are you're measuring a variable in the technique and not actural results.) Put another way, I'd like to understand this phemomenon and pattern. Someone out there with experience, must have a clue which sites are accurate, if any.

Also, as I said if I'm getting .8 or 1.3M and I'm supposed to be getting UP to 6M, then there's likely a problem on the line. That would make it MUCH easier to tell comcast about my intermitent TV problem and get taken seriously. Intermittent is a deathnell to a problem, but if you have a 2nd piece of evidence... But I don't want to comment on it to them if these numbers are meaningless. And I want to add phone, and I don't want to do that until the line is okay.

What do you mean by "the comcast line I see if 6 M bits per second" What comcast line do you see? Where?

Only the very last site in my ps measured 10 million bits (and isn't one of the recommended sites). The others measured 1 million bits. That's only 1M, not 10M!!!! Nothing to be so happy about anyway (although that's not what my concern's are.)

You're going for an easy emotional comment in the response and I understand why, but I'm focused on a valid question. I hope with this explanation, you can understand it better.

To reterate, each site's results are:
1. .7 to .9 M each time repeated, no other range of results (cnet)
2. 1.2 to 1.1 each time repeated, no other range of results (speedtest)
3. 1.3 to 1.1 each time repeated (speakeasy's)
4. 8M to 6M each time repeated, (speedguide's)
That makes the results site dependent, not moment dependent.

Thanks.

- Collapse -
"Only the very last site in my ps measured 10 million bits"
Jan 1, 2008 6:01AM PST

"Only the very last site in my ps measured 10 million bits (and isn't one of the recommended sites). The others measured 1 million bits. That's only 1M, not 10M!!!! Nothing to be so happy about anyway (although that's not what my concern's are.)"

Bingo. Your connection to the internet is over 10 million bits!!!

Congratulations.

-> I can see where you might be confused. As you mover further and further away from your test site the speed is going to drop. It's a simple matter is that as your get farther out you are on the "pipe" with everyone else.

If you are going to want 6 megabit from your house to all those sites then somehow you'll have to tell the rest of the world to get off the internet while you test it.

Bob

- Collapse -
Bob don't respond to me again!
Jan 1, 2008 6:27AM PST

You aren't reading my question and are using this as an excuse to call me dumb. You're not getting it. And you are wasting my time.

Are you the same person who responded to the only other cnet post I ever entered? You sound the same. Very insistent on giving a "stock" reply for whatever without a clue in your head. In this case that "people complain to much about bandwidth results" and in the other case "dual memory machine means you better put in two," when I was asking about model numbers. Can you even read? Are you high on something and not concentrating? Is this a little ego trip for you?

One UNKOWN site saying 10M DOES NOT MEAN I'M GETTING 10 M, when reputable ones are much lower at 1 & .8M !! Any thinking person would know that. Also, I clearly said I don't expect 6M so what you clearly pointed out is that - YOU DIDN'T BOTHER READING WHAT I WROTE.

Please do not EVER respond to my posts again. If you do - i will do what suits best. Ignore the h-ll out of you. You are too stupid for words, and too attacking (and insulting) to be gently responded to. Good-bye.

- Collapse -
One correction
Jan 1, 2008 6:46AM PST

Speakeasy is 10. So leaves same question. Cnet is .8 (<1M), speedtest 1.15. That's a bizzare level of difference in measurements. Who's accurate (in the ballpark)?

Sorry for the mistake. Got distracted by bob.

- Collapse -
Sorry but it's really easy.
Jan 1, 2008 9:04AM PST

You are measuring different connections. The highest speed is the one from your PC to a network that is very close in the internet world.

The further out you go, the lower the numbers get. I know it seems like they should all come in at the same speed but let's try this explanation. It's about how your machine is connected to the internet and how your ISP has a "pipe" to other places on the internet.

Let's look at this article -> http://en.wikipedia.org/wiki/Internet#Internet_structure

The internet is not just one pipe but a collection of networks. Your measurements will hop across networks and take a speed hit along the way depending on traffic and more.

Back to your measurements. Your cable connection is rated at 6 megabits which you have proven to be better than this. But since your ISP does not own the rest of the internet they will not be able to guarantee you 6 megabits from your machine to another machine that is not on their network.

I'm unsure why you are upset over this discussion. It is hard at first for some to get enough information to figure this out but I'll keep trying.

Bob

- Collapse -
Bob - Not sure why you're still talking about the pipeline?
Jan 2, 2008 9:56AM PST

You don't seem focused on the question yet... Think of it this way. I have four radar detectors on the dash. At very similar moments in time they show speeds respectively of 10 mph, 30 mph, 100 mph and 100 mph. Overtime you drive the same road and they each vary 5 mph from their original readings. So then which detector is in the ballpark?

All the tests I did required me to pick a server location and I picked a local one (I'm in a big area - so it's one on all the lists.) I also tried distant server locations and the results basically were the same per site. (That had surprised me.)

This last OOL test was maybe similar to 50mph, yet another number range.

If it was a matter of pipeline, of outside usage effects, causing variablity, then ALL the readings of EVERY SITE would be all over the place over the WHOLE range of possiblities. These are site specific variances, making my question - which site is accurate...if any? Is something else effecting my testing that each site is reacting to differently (such a firewall I don't know if running.)

If each of these test sites tests speed in a different way that results in consistently different ranges of numbers, then they are meaningless. Unless they include comparative ranges by which to judge their tests. (Blood tests include normal ranges because different methods results in different ranges and are meaningless as single point numbers.)

I think for now I won't say anything postive or neg about the broadband, and just ask comcast to looking to the TV blips, before or ask I sign up for phone service. My folks spent a very tough couple months getting their digital phone service to work (and their company was much more responsive than my past experiences with comcast), and they had a 2nd line back up - hence my concerns.

Thanks for the info. I take back the 'dumb' comment. It wasn't what I meant. I am frustrated because you seem to be repeating over and over some very, very basic facts (i.e. it's a shared network with multi nodes). And doing it even with some glee at being able to criticize someone for if they were questioning that. (That's the part that got me.) While missing that I wasn't asking that. I just want to know why all the radar detectors on the market aren't giving usuable results, when I see references to these same ones all over the place.
- C

- Collapse -
Yes, I think you are getting it.
Jan 2, 2008 10:37AM PST

You should see different numbers from each speed test. The HIGHEST number will be your internet connection speed. All the others are being affected by how fast the pipes are from you to that test location with more reductions due to other people using the pipes.

I think it's about to be very clear in your mind what is going on.

Bob

- Collapse -
Try this.
Jan 1, 2008 7:30PM PST

Click the link, download the file. At the half way mark check your download speed and cancel the download. This will give you a good indication of your download speed. This is the test that OOL users use.
ftp://ftp1.optonline.net/pub/test64

- Collapse -
Tried it!
Jan 2, 2008 9:26AM PST

Thanks Ed

At the halfway it was 380KB/sec. If capital B is bytes (not bits) then that's 3040 kbits/sec which is 3Mbps. Which I'll take as being a very decent rate for a 6 Mbps connection purchase (true?).

Also, I think I can safely assume that a community paying for higher speeds has developed a stronger sense of which testing methods/sites work well. So I think that answers the question! From the dslreports OOL page which linked to the test I wound up linking to a company Visualware near me that showed very similar results 258 & 320M. I still wonder how these well-known test sites can be so different and there's no obvious commentary out there about it?

Some of the other test numbers like pause and qos weren't that good, but I'll need to learn a lot more before I evalute those...

Meanwhile netgear's 824 is working very well. I may still try the belkin N and compare. (The belkin is 108 & longer warrentee.) For the phone, I'm going to rewire my jacks to two lines (I've done it before), put it on the 2nd, then port the number and cancel the first line AFTER it's working well (comcast said that's perfectly do-able.)

Thanks : ).

- Collapse -
I get 600+KB from a
Jan 2, 2008 12:01PM PST

6000kb DSL line. http://img98.imageshack.us/img98/6290/testlc3.jpg
So IMO, 380KB from a 6000kb cable line is not to good. But with cable being a shared line, depending where you live and when you test, that might be all you get. Have you gone this route? http://www.dslreports.com/forum/tweaks

As to the various speed tests, you have java, browser and flash tests. You have distance and routing to the test site, which plays a big part in your result. With my DSL, I always like ftp download tests and java tests. Java test coming from my ISP site, which is very close and gives a good indication of speed, up and down before or just after the net. If I test to a java California site, my speed drops in half. You can play with these sites http://www.dslreports.com/speedtest?more=1 just for fun if you want.
Ed

- Collapse -
Thanks some more!
Jan 5, 2008 5:19AM PST

That's very funny...the test site you linked to...is the same one that I got that local place from that I mentioned above! I have to try the tweaking forum ideas (after I test the two routers and settle on that.) At least I'm hooked into getting started now!!

I'm not sure I understand how to do a java test, but I should be able to figure it out. I wonder if there's a file on comcast's site itself I can download ftp and be more accurate? (I think they're not locally based though.)

I agree that 3M would be small for a 6M connect, but I'm in a high congestion area, and I saw other people's test results for comcast (that same site) at 3M too. I'll have to ask my neighbors and see what's "normal" here. My other numbers, latency, and such were off so there may well be some tweaking to do.

My catch is that I'm at 17460 of DSL's 18000 copper wire limit and no FIOS available. Because of the distance, Verizon very graciously could offer me up to 1.5M but at the 3M package price. So cable it is.

Since getting to the test site plays a part in the result, then the test isn't very meaningful. If a race started from when everyone left their house to get to the race site - instead of from the starting line, it wouldn't be a very useful race. There should be a tool to download that tests from your location to the exit point for your provider network. (I have no idea how to design one.)

Thanks again!!!!

- Collapse -
just to clarify one thing, please
Jan 1, 2008 7:58PM PST

you said "One UNKOWN site saying 10M DOES NOT MEAN I'M GETTING 10 M"

so what DOES it mean?

i would love for the pages on Cnet to appear at the speed of light, but it's a sad fact that sites in the UK will be asking me if i want coffee while Cnet is still asking me what i want for the first course

and to REALLY make you happy, using speedtest i get 1740 up and 150 down from the Washington site
Wink

jonah

.,

- Collapse -
Jonah - it means...
Jan 2, 2008 10:02AM PST

"10M" It means nothing. I explained why with radar detectors in my comment to Bob.

I haven't tried any sites over the water...I wonder if I'll get the same slowness you get from the servers here... Oh well.

I never said I was unhappy with my connection Happy - just with my TV blips (it's moved from white lines back to stoppages), and was arming myself. Also, had what I thought would be an easy question.

Cheerio,
C

- Collapse -
Those 10k measures are not useful
Jan 6, 2008 10:26PM PST

To add clarification

If my connection speed is 3K and the route to their server is 15k because others (they and others along the way) have invested in higher speeds than I have - the resullting speed given to me is the average of mine and their's. So, not meaningful. And can easly be higher than 6M.

I'm still stunned there's not time stamp in the test process so the time from my provider to their site can be discounted. There's also something strange when I get the same results from close servers on their lists as at far ones.

- Collapse -
Look at it like driving on a motorway, (freeway).
Jan 1, 2008 8:24PM PST

You seem to be upset at what Bob says, but what he is telling is the simple truth. Your broadband internet connection speed will not be constant, and will fluctuate wildly.

Here in the UK there was a TV program that started a campaign against Internet Service Providers, (ISP's), about their broadband speed, and the program had it's own broadband speed test and form for viewers/users to complete. In case it is of any interest the site is here;
http://gadgetshow.five.tv/jsp/speed_test.htm

The premise of their campaign was that ISP's sell us broadband connections with the words, "Up to xxx mbs", (mega bits per second), where xxx is the stated speed. The campaign wanted ISP's to be more truthful and say what actual speeds they could supply. That's all very well and the ISP's could change their wording, but what are they going to say? Here are some examples I would offer;

"Up to 6 mbs but only when there is no other traffic on the internet".
"Up to 6 mbs but only when your neighbors are not downloading music and videos off the internet when you are online".
"Between 1 mbs and 6 mbs, depending on the traffic on the internet at any second, (make that a millisecond)".
"Maximum of 6 mbs, but we cannot promise you will get this at any particular time of the day or night".

None of these, or anything else they could say, are, in my view, particularly attractive as advertisements, and "up to xxx mbs" tells me near enough what I need.

I don't know if you drive, but if so, consider driving on a motorway, (I suppose it would be a freeway in the US). If you're in the fast lane then you can drive as fast as you like. but if the vehicle in front of you is traveling slower than you and all other lanes are full, then you can do little but travel at the same speed. At busy times, commuter time of day or holiday times, the traffic is going to be busy. At other times it may not be, for example late at night, but it only takes one blockage to slow down traffic to a stop.

That's broadband internet. You might be surfing along at your maximum speed and then "crash", you're down to a crawl.

I look at my own broadband connection and I consider these things;

1] Is it on? Usually it is on immediately I start up my computer.

2] Can I surf the internet quickly? Usually yes. Sometimes no. But those times the cause could be many things.

3] Do I lose my connection at all? Usually no but sometimes yes, (nothing is perfect), but it is getting better. My ISP fixes it quickly and usually without any intervention from me, and such downtime is getting less and less.

4] Can I download files quickly? I can download music and videos quicker than I can listen/watch them. That's quick enough for me.

5] Do I want it faster? Of course, but I know the technology and infrastructure has to improve before that is likely to happen.

One other thing. Don't get confused with results that show Download and Upload speeds. Download is the data your computer receives and Upload is where you send files out from your computer. Generally download speeds will be 10 x faster than upload speeds, whatever ISP you use, and whatever type of connection.

I am not sure why you are having problems accepting the information you have been given.

Mark

- Collapse -
Mark - you also didn't understand my question
Jan 2, 2008 10:11AM PST

To clarify " Your broadband internet connection speed will not be constant, and will fluctuate wildly."

DUH!!! Also, I wasn't confusing download and upload. Although actually I haven't hit a slowdown yet of any noticable amount (in several days time and lots of logged hours online.)

My question was a lot more sophisticated than those who don't understand shared pipelines, or the crafty meaning of "up to".

I too don't understand, after all this discussion, why you too would be so simplistic in your answer and not read the question itself. I found you "why can't you accept" comment insulting. Why can't you accept that someone might be asking something different for a change?

Lotta luck,
C

- Collapse -
The RADAR GUN ANALOGY.
Jan 2, 2008 11:01PM PST

I like that one. But remember you are not measuring your PC's (car) speed, but other machine connection speeds.

It's not unlike you pulling up in your car to some intersection and pointing the gun in each direction. You'll get different results in each lane and with each car.

What a good analogy.

Bob

- Collapse -
Hum,
Jan 5, 2008 5:38AM PST

Thanks. I liked that image too.

Your analogy doesn't quit work. It's more like there's a bunch of guns each designed with different filters and transmitters so they come up with very different speeds for the same motion (same car at same speed).

Or maybe to use yours it's as though all the other cars are being measured, but they're reporting it as the speed of your car.

Making the answer to my question, not that the internet is varied pipeline activity (which I knew), but that these tests are close to meaningless, and these test creators haven't done a blessed thing to mitigate for it. I assumed that if everyone was recommending these test sites, there was something in the design to make them speed tests, not random tests. Given the consistency of the test results with only 5% varance per site, the variablity is something fixed that could be worked with, which makes this all the more particularly poor quality. I assumed some forumla or formulation in the test process.

I wasn't having trouble understanding concepts that way you kept implying. I was having trouble understanding an old problem in humanity - that things could be as "poor quality" (substitute other words as desired) as they are - and no one mentions that important fact. It's the one thing I missed the most when I left IBM - as funky as the company could be (actually mostly a good place) - the people were smart and if I'd said something like this - it would have been instantly understood, and someone would have worked out a solution already. When I left for other companies the lack of that baffled me for a while until I got more used to it. No one would have sounded insulting to me because they would have seen it with the same expection I did and understood my frustration.

Oh well, on to other things. This all beats the horse and buggy dial up I was using before! No radar detector needed for that!!!!

- Collapse -
Download Speed
Feb 24, 2010 1:31AM PST

Hello all,
I have a stupid router HG556a (Gateway).
When I enter the router and check the service status I get:

You are connected.
Your speed is 6 Mb/s.
Connected to the Internet via ADSL

But when I run any test speed from different website I barely get 1Mb/s
I run Window Vista Ultimate built by myself with 3.6ghz processor and 4MB ram.
I used the same computer when I was living in New York and I was downloading up to 25Mb. Right now I live in Barcelona and the most speed I could contract is 6Mb/s but again I barely get 1M.
Actually during the morning goes up to 3M or 4M but at this time (6:36PM) and later on in the evening goes down.
Again my router tells me:
You are connected.

Your speed is 6 Mb/s.

Connected to the Internet via ADSL

Any help to make this download faster
Thanks in advance
Emanuele

- Collapse -
Broadband - Speed test
Jun 22, 2010 4:34AM PDT

Hi I am sure that you are a fed up as the rest of us.We are all hoping for good speed and it's hard to prove what we are getting. Ping Speed Tests are not very good. The do it once a day and the results do not tell us why our connection is slow. We created WebMeter which tells you all day every day what speed what you are getting and why. Its tells you which websites are falling down. It also tells you if you are troubled with other problems. Ultimately it tells you what you want and need to know. It kicks the *** of all other speed tests!!!