Web Development

Ask something
Discussion Locked

Bandwidth controversy

0 helpful
13 replies

by mtrox 1 view

Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

Not sure if this is the right discussion for this one but would love some help. I wrote an article in a local business journal on small business networking. In one paragraph I talked about how speed on one end doesn?t mean anything if the speed on the other end is slower. I gave the example of one of the local ISPs who is now offering 5 Megs of down speed, and made the point that the average guy will never see that speed since there are no sites I know of that upload at that speed.

Now the ISP wants a retraction because they tell me web sites routinely push data out at 10 to 44 Megs. I know that but that?s not my point. I told them, if there is a time when only two people are connected to a web server with a 10 Meg connection, then on that rare occasion those two users could actually see 5 Megs a piece. If that happens routinely, you need some more use for that server and it?s bandwidth.

If I had told web site builders that no one is uploading at 5 Megs I?d be dead wrong. But I was talking to users and my point was that the bandwidth gets split among all users attached at a given moment so that no one gets anywhere near the full speed.

Am I wrong? If I?m right, anyone now of a site that will back me up? I Googled for a while but only found one that really talked about this issue.

13 replies

by oz_media

Ever since I started with 5 Mbps connection i have realized just how many sites DO offer these speeds.

TO elaborate on your example, and though it is a different scenario, a GAMING site offers a 660 MB download. I will get a STEADY 650KBps+ and there must be HUNDREDS of others doing the same thing at the same time. BUT that's a site designed for highbandwidth downloads.

When sending large music files to companies in Emgland or Germany (where partner studios operate) they also have at LEAST 5 meg pipes. But even then I get a nice steady 650KBps for the most part.

You DO have a point, IF the other guy only has 1MB throughput, you will NOT get your 5 megs down from him, but MOST business class services will offer bandwidth routing and you will get nearly peak speeds if they have a similar sized pipe.

I would have to read your article (in fact I would be quite interested) to perhaps determine what WORDING the objected to though.

You DO have a point, but it's based more on a variable.

It's no different than saying that residential shared broadband MAY differ in speeds based on the number of users online sharing your port. But it doesn't make a case for the lack of service as much as it explains a POSSIBLE varible that COULD reflect on speeds. OR tht aa car can do 80 miles per hour, but if someone is driving slow in the fast lane you only get 50 miles per hour. It doesn't reflect on the car too well and it is a variable for possible consideration, not a constant factor.

Perhaps there was something else in your article they objected to and just used this as the simplest reason.

by jck

I see your point. I don't want to get into corporate names in a public forum, but I just signed on with a cable company because they are advertising 5Mb/s download, 384kb/s upload speeds. Well, I tested it right in front of the install technician. I could get speed results in that range *within* their internal network. However when I went to route out onto the backbone to any other site outside of their internal network, their speed immediately decreased (the fastest speeds I could find) to around 3.1Mb/s.

It's still fast, but it's not a blanket speed for all connections on the internet as they'd let you assume. So in a way, they (as a business) would understandably be upset that you are putting somewhat of a negative light on their service. However, they should also take responsibility for stating that they offer "5 meg download speed", when it is clear those who are technically aware that it should carry the caveat "...from accessible sites offering sufficient bandwidth within networks under our control" instead of that innocent little preamble "up to...".

I wonder if I could sell a service putting 100Mb/s hubs/routers in peoples' homes and advertise that "I offer up to 100Mb/s speeds! Call today!"?

I know...absurd...but in a sense, that's what they're doing...just with better PR and lawyers.

by oz_media

The both services I use clearly advertise UP TO 5Megs. I 'average' 4.6-4.95 Megs through almost ALL test sites, local, US or overseas.

I get a solid 650KBps from MOST sites but the lowest has been around 350 KBps, which is obviously a 3 MEG pipe. But that's still DOUBLE basic DSL here and the difference between 3 meg cable and 5 is only $10.00/mo.

by jck

Well, I live about 45 minutes outside of a metro area of 2.5M folks. Hence, I might not get the best performance because:

a) I'm in the "boonies"
b) My cable installer works as a "toothpick engineer" at his primary job.
c) gremlins inhabit my Netgear 802.11g wireless router.

All silliness aside, I did note that they use the "up to" qualifier with their ads. Just that I think it a way to cop-out when their network only runs optimally at about 2/3 of that maximum, as is the case with my service at home.

I see claims of the same broadband carrier I have in other large metro areas reaching 8,000Kb/s download speed, at which time I drool on my desk and dream of how much faster I could download and game online.

In reality, they probably didn't run the best copper here and I probably have a slew of 12-16 year olds on my branch of the network hitting Kazaa and WinMX 24/7. And, I've heard some cities have fiber up to the residence now. I know for a fact here it is all copper to the junction box (because I watched them run the new coax when I bought my house and got cable and internet).

I wish I got 650KB/s. Most I think I've seen is 380-440KB/s. Most of mine run 120-270KB/s. Even Microsoft runs no more than about 230KB/s.

Oh well. The hazards of living in "the sticks". How much is a DS3 again?

by tomsal

When it comes to the networking world, its a rarity that the posted speed of a connection, any connection, is the true speed.

You may think right now that you are talking to your LAN at 10 megs because you have a 10 meg card, but you aren't. You may think your home's wireless network is running at 54, but you aren't.

Like many things in life, there is overhead in networking as well. Add to that we are talking electrical signals (light if fiber, radio if wireless) that are not 100% in the same perfect state all the time. Variances in current, even atmospheric conditions (esp. for wireless), then the common effects from NEXT, EMI, the physical media (distance/condition), etc.

All this stuff on top of the overhead of the protocols themselves is what robs from the "posted" speed of a connection. This is why it is nearly IMPOSSIBLE to truly get a consistent speed as advertised like 10 mb/100mb, etc.

That all said...I think I know the company you are talking about as I believe I use them also for my broadband. lol.

PS. my connection at home is great nearly all the time.

by jck

Well, it is true that nothing is 100% of optimal performance in electronics. That is a given. But I'm sure you'd agree that the speed degrade of your PC across a 802.11g network within your home wouldn't suffer excessively without some abnormal condition occuring. And, the overhead of protocol layering, packetization, etc., would not cause a major impact to your network performance, i.e.- nowhere near a 25% loss of normal operating speed. And, the likelihood of interference across cat-5 or fiber is even less likely than that of RF interference in a wireless environment.

With that said, I just see it as strange for them to claim that someone is misrepresenting their internet performance and demand a retraction, when most likely their daily advertising makes a claim of a maximum speed that most users will never even get near in day-to-day web page retrieval and file downloads.

I know that my average download speed is around 2-3Mb/s when doing file transfers from most large-name websites who do carry large bandwidth connections. Although that is still "high-speed", I've yet to see 5Mb/s from a source outside of my ISP's network and rarely attain 4Mb/s. The reason I stay with it is because it is a better deal than DSL here, not necessarily because of better performance.

Simply put, their sales technique does prey on the uneducated and uninformed. Is that planned or part of their strategy? Who knows. I can't and won't speculate.

In my opinion: if they expect technical individuals not to speak negatively of their product and their claims of it's performance, then I see it as their responsibility to either improve its performance to a level which eliminates said negative claims, or to modify their advertised claims which are in question to give a more realistic assessment of average performance.

Not to say I totally defend the original statement for which a retraction was requested, since I haven't seen the original text. But, it's almost certain in the modern age of advertising that they've used that good old line "...up to 70 times faster than dialup, and 3 times faster than conventional DSL." which leads a lot of customers to assume and believe they will get 3 times their DSL and sign up right away.

Also...my connection at home has been pretty good as well, with exception of 1 overnight interruption (7 hours before I went to bed), and 3 service interruptions to my office (we use them as our high-speed fiber carrier at my job) which were all issues with their internal network.

With that said, I think there's something to be said for companies to exercise "total disclosure".

by mtrox

Appreciate the feedback. I?ve seen what some of you guys have seen. I?m on a Comcast Cable modem. I think they advertise 4 Megs right now and I routinely test at about 3,500-3,700 Kb/sec. If I download from Microsoft I see transfer rates of about 370 KB/sec, or a darn fast 3 Megs.

My comments were in a paragraph talking about how speed claims are not always meaningful. I talked about this 5 Meg claim?.and that for most people, 802.11b on a bad day is running faster than their internet connection ever will. Since I want this column to read like something the average guy (i.e. non-geek type) would read, I don?t have the space to put in a lot of qualifiers. This ISP wrote me a polite letter of protest. That?s fair. And although technically they?re right, web sites do upload at 10 to 44 Megs, for the average guy who?s making a t-time while his daughter IM?s with her friends, 512 Kb is about the same as 5 Megs in the real world. And at 5 Megs, the limiting factor on your internet connection is not likely to be the size of your pipe.

I?ve never tested this but I?m on different computers every day and my experience has been that it appears that once you get a computer all alone on a connection of at least 512 Kb, the real difference in the speed of loading a page is the computer. How good is the video card? How high is the video set? What are the MTU and TCP receive window settings? Blah, blah, blah.

I think these guys got mad and got over it. But I?m going to do a test. I?m going to take my old Dell notebook over to a buddy?s 768 DSL, a couple of this ISP?s ?5 Meg? cable modems, and a few other internet connections. I?ll delete all temp files, run a couple speed tests to benchmark the speed, then go to the same 5 or 6 sites and see how long it takes a page to fill. For the average surfer, that is the bottom line, how long does it take to download and paint a page. Is that a fair test? Anyone seen a test like that?

by jck

The only suggestions I can make to your testing are:

The repaint time within the GUI environment has really nothing to do with the download speed of the network. Timing the period from click to the draw of the screen will give a ballpark of speed, but it won't give an accurate assessment of the true transfer rate of data. And if you are using an older computer to surf, then the display of the page becomes more increasingly dependent on how fast the computer can load and display all the graphics, rather than how fast the pipeline provides the data.

But since your assessment is to provide people/businesses with information on what site ISP has best throughput, the most fair comparison to all parties would be to setup a computer and set it to retrieve the same sites and time them during the same period of the day and perhaps even do multiple retrieves and give the max, min, and average times.

Does sound like they got over it. At least you don't have some attorney hovering over you with threat of tort. Good luck with that. Sounds like you are at least making the best attempt to give the readers fair assessments.

by mtrox

Ick, I agree with what you said and if I had more room in my column, I might have said all that. My entire point was...or at least I meant it to be, that 5 Megs doesn't mean anything unless every other link in the chain can keep up with that speed. As you say, the time to paint a new screen is VERY dependent on the computer...just as the download time is VERY dependent on the upload speed at the other end.

I just gave an opinion. I don't think they would take the time to get lawyers involved. I just want to be fair. To be honest, I am now on another cable modem ISP and wish I was still on the BeepBeep one. I don't have it in for them at all. Wish I could go back to them.

by oz_media

There is one that just loads thousands of company logos for a predetermined test time and then tells you your bandwidth and how many thousand images displayed. A good inidcator of web page load speed. Pretty cool actually! But I don't have the URL handy.

Then there's broadbandreports, my fav as it has a really good tweak test and recommendations, DRTCP etc. SpeakEasy servers are not really teh best for me due to location but it still flies for me.

by oz_media

I just realized and considered another fact about the 5 meg limit that I have forgotten to mention.

I am currently downloading a somg from a server noted at a cap of 1000Mbps so I am getting about 318KB avg (which is still way over the servers cap isn't it)
BUT, at the same time, I am downloading a demo at 328KBps form another site at the same time.

So the added speed IS really beneficial if you require simultaneous downloads at high speed. Whereas before (3Meg) I would get 175 or so for downloads now i get some at over 400 while at the same time getting another one or two if not more at high speed also.

So for people with multiple file download needs, they can get more files simultaneously than with a slower connection. It really does save time.

by jaqui

that takes time even on a high speed connection.

download a full 3 cd distro from a public ftp server.

my personal best download speed was 1 hour total elasped time from click to downloaded and saved.
695, 699, 695 mb iso files.

no download accelerators, not using bittorent or jigdo.

just straight browser ftp download, under a gui
( mozilla suite under kde )

avarage on the 3 mb connection is 5 hours.

by oz_media

And 5 Megs is lightning quick even when compared to the 3Meg pipe.

I find 5 Meg cable in the mainland is a fair bit slower than at home but for the most part it is still okay.

My Sat, which I usually use on when over there is pretty consistent though, good thing I don't get billed for it, it is REALLY expensive!

This discussion is closed to new replies.