Comcast and Net Speed Tests

JimDaGeek writes "I recently moved to Columbia, SC where I have Time Warner as my cable ISP and pay for an 8 Mbps connection and have been very happy with the service, speed, and reliability. In contrast I have heard bad things about Comcast. So now that I am up in the Philadelphia PA area visiting my parents, I decided to test out the speed and reliability using the Speakeasy speed test. The results surprised me. Here are the reported download speeds in Kbps: New York, 18,946; Washington, 15,821; Atlanta, 11,257; Chicago, 10,042; San Francisco, 4,230. What is going on? I know my father is not paying for a 10+ Mbps connection. Is Comcast giving priority to popular speed-test sites?" From Comcast's site, in the Philadelphia area they seem to offer download speeds of 6 or 8 Mbps, with an option for a "PowerBoost" to 12 Mbps on large files. This wouldn't explain the results JimDaGeek got of almost 19 Mbps down.

Update: 07/10 12:07 GMT by KD : A friend in Massachusetts had a tree fall on his house. The Comcast guy who reconnected the lines told him that they are boosting the line speed to 20 Mbps down / 2 Mbps up in certain areas to be more competitive with Verizon FiOS.

Comments

Sounds like Comcast Blast! test area

[Anonymous Coward]

Some parts of Philadelphia get 16/2 as part of a new service test, so the New York test is a bit high, but not unheard of for cable. I just tested my speeds, and nothing unusual.

Re:Sounds like Comcast Blast! test area

[GizmoToy]

That's a great transfer rate. Do you happen to know what they're charging for that speed in Philadelphia? Nothing affordable even comes close around here.

Re:Sounds like Comcast Blast! test area

[RMingin]

I'm paying $55/month for their Speeed Tier service. Standard service was 6/768 and Speed Tier was 8/1Mbps when I worked there briefly, they periodically give small increments to the speed. Speed testing websites aside, I always hit GigaNews at a very solid 3MB/s (MB not Mb) for as many GB as I want, and my CS:S pings are often single digit. I'm fairly happy with my service.

SpeedTest.net

[ragnarok]

The Speakeasy speed test is just a re-branded version of speedtest.net. They have a lot more test locations to choose from there.

Re:SpeedTest.net

[Lumpy]

and it's a worthless test.

One big transfer != bandwidth capabilities.

Give me 100-500 smaller files with smaller ACK going back. that gives you a real test that will show latency and jitter.

also check many different ports. Port 80 get's priority. Ports above 8000 get lower priority. Ports for Voip are screwed with hard.

Re:SpeedTest.net

[timeOday]

Ports for Voip are screwed with hard.

I'm on Comcast and dropped my home line for Vonage about two years ago, and in my case it works very well. I don't like Comcast's pricing, since Cable+Internet alone run $100/mo. Ouch. And sometimes I'd like a higher upstream bandwidth cap. But the fact is (for me) the high-speed internet works pretty darn well, both reliability and speed.

Re:SpeedTest.net

[eli pabst]

Give me 100-500 smaller files with smaller ACK going back. that gives you a real test that will show latency and jitter. For the average joe the main content where bandwidth is going to be an issue are things like an html file, audio file, video file... all effectively one large file. They don't care about jitter, just that their homepage loads 10x faster. Obviously it's not a complete metric, but for most people it's a reasonable measure.

Re:SpeedTest.net

[Anonymous Coward]

bandwidth != latency. You want both (high bandwidth and low latency, that is), but that doesn't make them equal.

Re:SpeedTest.net

[RegularFry]

Someone mod this up, please... bandwidth is a rigidly defined term that the GP doesn't understand. A large file download is a perfectly valid way to estimate it. Of course, it might not be a realistic measure of how usable a connection is, but that's a different matter.

Re:SpeedTest.net

[nberardi]

Comcast's headquarters are in Philly. Philly usually gets better speeds just because of the short physical distance between the main servers and the home. Plus if Comcast is looking at rolling out a new feature such as 20 Mbps they usually do this in Philly first. I used Comcast for many years in the Philly area, and when they upped the speeds to 6 Mbps back in 2004 I was one of the first to see it in the nation even though they were still only advertising around 2 Mbps.

Re:SpeedTest.net

[Anonymous Coward]

Right, jitter may not be much of a problem for the average joe, since he's probably not using voip. However, the average joe would certainly be better served with a speed test that does send many smaller files, as the parent poster mentioned, since very few HTML pages are comprised of a single file, but rather references to dozens, sometimes hundreds, of separate files including images, javascript, etc.

Re:SpeedTest.net

[charleste]

Maybe it depends on where you are? I get an accurate 8MB without my work VPN, and a typically slow 260ish KB when I have my work VPN on.

Re:SpeedTest.net

[Pharmboy]

For the average joe the main content where bandwidth is going to be an issue are things like an html file, audio file, video file... all effectively one large file.

html isn't one big file, just like bittorrent files aren't (are hacked into smaller chucks). Most video and audio are streaming, which uses UDP instead of TCP, so you get different "bandwidth" results, since measuring with TCP would account for dropped or corrupt packets, and UDP just ignores them. I understand your logic, but your example is a bit flawed. The point is there are several ways to measure an internet connection, and bandwidth is just one.

Re:SpeedTest.net

[fm6]

In other words, the SpeedTest benchmark is what engineers call a "microbenchmark". Meaning it measure some trivial function that isn't usually a key bottleneck.

OK then, how about a benchmark that does measure a real-world use case? Anyone?

Re:SpeedTest.net

[Anonymous Coward]

Good Old Comcast ... We were getting 5 to 6 Mbs on long DL's. But about 4 months ago they changed nodes and now we get 4 at best and a lot of times it's just 2-3 Mbs. Pretty disappointing.. :(

It looks like they are capping the thru put and capping differently for different packet types or ports?? nntps caps at 2.9; https caps at 4 but last nite gave 5 Mbs But it was pretty intermittent (looks choppy on the net meter)

Am wondering about upgrading my Modem (Motorolla 4100) to Motorolla 5120 would that help?

Re:SpeedTest.net

[skidv]

A good test product is ttcp. TTCP measures both bandwidth and latency. You can learn more at http://www.pcausa.com/Utilities/pcattcp.htm.

You'll need a partner somewhere on the 'Net, but you'll get an accurate measurement.

The problem I always had with speedtest.net is that the data it transfers is static, so that content caching services can skew the validity of the data. Also, your own internal browser can cache the information so subsequent tests to the same website may reflect incorrect results.

Re:SpeedTest.net

[Bake]

Most video and audio are streaming, which uses UDP instead of TCP

I keep hearing this over and over again, but I've never actually come across a site in the wild that actually does that. What common streaming protocols actually use UDP instead of TCP? And what popular sites actually USE them?

Re:SpeedTest.net

[Pharmboy]

I can't speak for video, but the gaming system Steam uses UDP for all the game play, and tcp only to log into the server, so 99% of the packets are UDP. Any stream that is live is UDP, ie: if it misses a packet, you just don't get it. This would be for sports video on the web, etc. Even DNS uses UDP. I don't sniff packets enough to see who does and doesn't but I know that it takes a lot less overhead and bandwidth since there are no ACKS, only occasional "i am still here" tcp packets.

Re:SpeedTest.net

[Achromatic1978]

Yes, it'd help, the 4100 only has 10mbps, IIRC, and a slower CPU... I know that Telstra in Australia gives you a 51xx series cable modem when you use or upgrade to their "extreme" service (17mbps/256kbps).

Re:SpeedTest.net

[Anonymous Coward]

In Soviet Russia, CommieCast download you!

Time of day?

[carlivar]

Since cable bandwidth is shared, wouldn't the time of the test matter? I've noted (very unscientifically) that my Internet seems slower between roughly 7-9pm (on Charter in Los Angeles area).

Re:Time of day?

[Aeron65432]

Very true. My university has a dual-T3 connection, and it's fast as hell. Except around 6-8pm, when everyone seems to be on. You'd be surprised at how much everyone on AIM and browsing the internet kills your speed, despite their low-bandwidth usage.

Then again, my torrenting habits probably don't help.

Re:Time of day?

[catwh0re]

It also makes me think of how complicated the simple "bandwidth test" will be should net neutrality get thrown out the window.

Already we juggle the factors of location, "paid for speed", shared bandwidth issues with daytime or peak traffic.. but then without some kind of neutrality we'd also be juggling whether or not the interconnects between yourself and the test site are all on a higher priority or lower priority pipe.(something we could never know)

Today your ISP can blame a bit of the slowdowns on network conditions, but ultimately it's obvious if your ISP is a slower provider.. but in the future they'll be able to knowingly serve you slow speeds while claiming it's just the low-priority sites you may be visiting.

Re:Time of day?

[kaufmanmoore]

Facebook and myspace can eat bandwith with all the crap some people post.

Re:Time of day?

[lancejjj]

Since cable bandwidth is shared, wouldn't the time of the test matter? FYI, practically all internet traffic traverses shared communication lines. The point-to-point connection of DSL becomes shared as soon as your local copper pair is squeezed onto a multiplexed line, within a few thousand feet of your home.

Re:Time of day?

[MrShaggy]

Since cable bandwidth is shared, wouldn't the time of the test matter? I've noted (very unscientifically) that my Internet seems slower between roughly 7-9pm (on Charter in Los Angeles area). I look at the difference between cable and dsl as 2 different ways of getting to a train-station.


Cable might be like taking a bus to the station. There might be other people there. However, its not that significant.


Dsl is like taking your car or a cab to the station. It definitely is not a shared ride into the station.


The one thing that happens is that everyone gets off of there rides, and they all take the train out. Pretty much at the same speed.

Of course during that time of the day there are many people on the internet, so it won't matter what you are using. Its somewhat insignificant how 'slow' it is. The only thing that kills the dsl is how far away you are from the office. Thats pretty wild. The thing is that you can get speed tests from eith side of the country or even planet. So even if it isnt the rush hour here, it will be elsewhere.


Thats why I think most dsl ads are so misleading.

Re:Time of day?

[Firehed]

You remind me of the time that I single-handedly saturated half the pipe for the entire school with a couple of unusually fast torrents. Oh, what I wouldn't do for a personal T3...

Re:Time of day?

[dbitter1]

Cable might be like taking a bus to the station. There might be other people there. However, its not that significant. Unless your bus is, say, giving away free money/crack/heroin/whatever and starting in a high-density housing project. Then it would be a ride like my former.

In Chicago, Comcast acquired their monopoly from several other smaller cable companies. The one I had apparently had some pretty sleazy workings- the cable modem appeared to be in some sort of bridge mode, as I could see ARP requests from other modems. For one weekend, I counted over two thousand unique MAC addresses worth of ARP requests.

And the type of speeds you'd expect from sharing that pipe with two thousand other people- barely above ISDN.

Re:Time of day?

[Matey-O]

Oh, what I wouldn't do for a personal T3...

What, short of paying money for it?

Re:Time of day?

[bluebonics]

this is because even if programs aren't using bandwidth networks use a csma/cd "carrier sense multiple access with collision detection" setup where, simply put, if someone is transmitting on the line, you back off and wait to transmit. so when 6-8 comes around and everyone starts on aim and myspace the network gets, for lack of a better term, congested with their packets; the tubes get clogged.

Re:Time of day?

[ls671]

It took more time to cable companies and costed them more money to adapt and provide good speeds during peek times due to the shared pipe between several customers problem.

Why? Because they had to multiply the number of hubs out in the field so the shared pipe would only be shared between a smaller number of customers (maybe a few blocks in a city) thus guaranteeing to respect the minimum average speed advertised. From the hub in the field, it's fiber all the way to the gateway.

Nowadays, both DSL and Cable are pretty much on par except DSL only works close from the Telco switch because the Telcos never installed that many hubs out in the field ! On the other hand, cable needed hubs in the field to avoid congestion.

As far as getting faster speeds than advertised, the speed you get is decided when you boot up your cable-modem. The modem downloads a config file by TFTP from your provider and it tells it it's maximum speed. If, for any reason they have problems with their configuration file system, they might default all customers to a given higher speed. Mistakes happends too in those systems and management doesn't put the focus on it because they just have got no way to tell that it's happening beside having customers on Torrent 24/24 and analysing their bandwidth usage if they can, their system isn't always as properly configured as we would normally imagine ;-)

Re:Time of day?

[StikyPad]

The internet is not a truck!

Re:Time of day?

[Anonymous Coward]

we'd also be juggling whether or not the interconnects between yourself and the test site are all on a higher priority or lower priority pipe.(something we could never know)

I can tell you that the bandwidth test sites will all get high-priority.

Re:Time of day?

[timeOday]

I've noted (very unscientifically) that my Internet seems slower between roughly 7-9pm (on Charter in Los Angeles area).

All the servers are more busy then, too, so who knows?

Re:Time of day?

[ivan256]

Yes, and your cable connection uses a shared upstream connection once it reaches the CO too. The difference is that the "last-mile" connection is also shared. Either can become a bottleneck depending on traffic.

Of course the internet traverses shared lines. That's practically the point of the internet.

Re:Time of day?

[Short Circuit]

That's nothing. I was in a vacant computer lab in Michigan during Spring Break, pulling Debian CD images from an FTP server in Germany (best transfer rate I could get; Jigdo never worked for me, and there weren't official torrent trackers at the time. {Are there now? Switched to Ubuntu a couple years back.})

The door to the lab bursts open, and a couple techs started methodically checking port labels on all 69 of the PCs, then started checking the Macs. (I was on a Mac 'cause those were the only machines that had burners.) They let me finish downloading disc 12, but told me not to start disc 13.

In 2003, it was kinda noticeable when someone had three simultaneous 200KB/s FTP transfers running. Fortunately, I wasn't in violation of the Acceptable Use Agreement, as I hadn't installed any unauthorized software...

Re:Time of day?

[Short Circuit]

this is because even if programs aren't using bandwidth networks use a csma/cd "carrier sense multiple access with collision detection" setup where, simply put, if someone is transmitting on the line, you back off and wait to transmit. While that's technically in the Ethernet standard, you don't see it much--if at all--with switched networks, and even less with networks with decent switches that hold packets in a queue. Switches and routers aim to eliminate that kind of inefficiency.

Re:Time of day?

[Short Circuit]

Yes and no. When your phone line gets multiplexed into a T1's stream, your line get a guaranteed throughput of 64Kb/s, at least within the context of that T1. What your voice line can actually transmit and receive is limited by other factors, though you can get the whole 64Kb/s if you spring for a digital BRI, which can be useful for things like ISDN.

Obviously, DSL isn't limited to 64Kb/s. The DSL signal gets intercepted and filtered out before your voice line gets to the T1, and the DSL traffic gets dropped onto its own packet-switched network at the DSLAM. And, yes, that packet-switched network may be shared among multiple customers, but in an asynchronous fashion. (That is, after all, a major reason packet-switched networks are so popular; They make it relatively easy to adapt to different load requirements.)

Re:Time of day?

[Deliveranc3]

Obvious vs proof, I live in Canada I've noticed my ISP (Rogers) limits bit torrent (specifically uploading to a single peer is 40-70KBps while multiple peers get 2-10KBPS)... How would I prove that?

There is already significant throttling going on, why wouldn't they prioritize some traffic very highly, simply cache the site and show your last mile bandwidth.

Totally underhanded, but these are fooking telecos

Re:Time of day?

[QuietLagoon]

The point-to-point connection of DSL becomes shared as soon as your local copper pair is squeezed onto a multiplexed line, within a few thousand feet of your home.

True, however, it is always easier (and less expensive) to add additional capacity in a central office than it is for a cable company to add more lines within a bottlenecked neighborhood.

Re:Time of day?

[livewire98801]

Technically he's correct. Cable by it's very nature is half-duplex. That's why we quit using thick/thinnet on LANs. Yet Comcast pitches it as the best internet connection available.

Thats why DSL and Fiber (okay, that one is obvious) are always better technologies than cable for last mile, no matter how much bandwidth you throw at it. Unless you're Verizon, and don't add another T to the DSLAM until you have a higher churn rate :-/

Re:Time of day?

[livewire98801]

Wonder if there's a way to spoof the TFTP source. . . or make it download from the LAN interface instead.

Yes, I'm a schemer. This is also academic, since I'm on DSL :)

Re:Time of day?

[node 3]

Technically he's correct. Cable by it's very nature is half-duplex. Is that true? I have always assumed cable internet used different frequencies for upstream vs downstream.

Thats why DSL and Fiber (okay, that one is obvious) are always better technologies than cable for last mile, no matter how much bandwidth you throw at it. What do you mean by "better"? DSL speeds are significantly slower than than cable speeds in my area, and cable is only slightly slower than FiOS (download speed, at least). I suppose you might have meant theoretically better, but you said "always", which is certainly not the case.

Re:Time of day?

[ls671]

Yes it is possible in some cases, mainly if their system and programming logic is poorly implemented. The best example that comes to my mind is the PAP-2 that Vonage sold for 10$ in some stores taking for granted one would have to use their service once they bought it.

Since it is legal to reprogram them because I owned the devices and I never used Vonage or sign anything with them anyway, I reconfigured a few to talk to our asterisk server. Re-configuration involved having an entry in my DNS server for vonage tftp server that pointed to my own tftp server in realty. So you could call that spoofing the tftp sources.

I have never tried to fool around with cable companies cable-modems although. But here is a hint: a cable-modem has usually two IPs; a public internet IP and a provider internal IP (like 10.10.10.33). The modem downloads its tftp file through the internal provider network. As an example, the tftp server ip might be 10.10.10.99.

Re:Time of day?

[CopaceticOpus]

What, short of paying money for it?

Dude, a T3 would be almost as expensive as just buying all those CDs and movies.

:)

Re:Time of day?

[MightyYar]

That may be technically true, but as a practical matter I've never seen my DSL speeds fall off of the peak, whereas my old cable service fell off of its peak every evening.

Of course, my DSL service is not as fast as my old cable service was...

Re:Time of day?

[BuckaBooBob]

Umm.. DSL is shared aswell.. Just at a different point.. It really depends on the engineering of your high speed as to which bottleneck is the first one to degrade your date rates. Where I live cable is 2-3 times faster than "dedicated" DSL no matter the time of day..

Re:Time of day?

[volkris]

There IS no neutrality, that's one of the myths you pro-NN people keep slinging about.

It's not a matter of "keeping the internet neutral"; ISPs do and have always shaped their traffic in various ways. Other threads under this story make present plenty of examples of this very thing.

So strike preserving a wondrous status quo from the reasons to support network neutrality legislation, and then ask yourself why such a blatant lie has been pushed by the corporations lobbying to have such regulation passed.

Re:Time of day?

[livewire98801]

I've had the Comcast "8mb service", and I never saw anything above 4mbps, and it was rare to get above 3. I have Speakeasy 6mb service now, and have only seen less than 5mb once.

Perhaps I threw "always" around a little casually, but that HAS been my experience.

Re:Time of day?

[Anonymous Coward]

Cable bandwidth is shared, as is the ENTIRE INTERNET. Why do people continue to emphasize sharing when it comes to cable modems or DSL?
You share EVERYTHING. Remember that lousy T1 to the office? That's shared. When you stay at an average hotel, they have a T1 connected to their wireless router. THAT is shared.

Cable modem, you're sharing a DS3 or an OC3, or sometimes an OC12 or GigE to your neighborhood. If you're out in the sticks served by DSL, you might have 4 to 8 T1's for your connection. Yes, it's shared.

The thing about Comcast (and just about every other provider) is that they are desparately trying to upsell you the big bandwidth connections, while at the same time, they're desparately trying to prevent you from using that bandwidth all the time. They want to control you so that you don't cause problems, so they don't have to spend more money providing a larger pipe to your neighborhood. They will say that 15% of the population ruin it for 85% of "typical" users. If you're in the 15% who run a bittorrent client 24 hours a day, and log into your home system from work and are pi$$ed because you're not pegging an 8 meg download 24x7, you're ranting and raving on Craigslist and /. Comcast says customers like those are the problem. The real problem is the expectation of "business" class service at a residential pricing.
And people complain when they have to spend $44.95 to $59.95 for this service?

Re:Time of day?

[Short Circuit]

The poster I was replying to was replying to a comment set in the context of a university network, not your run-of-the-monopoly cable provider.

I suppose the university might have been using DOCSYS for their dorms.

Re:Time of day?

[jaysones]

Yeah but if you got a T3 you could find Sarah Conner in just a matter of days.

Re:Time of day?

[Woundweavr]

No it is not!

Its a series of tubes.

Re:Time of day?

[Mitsoid]

Yup, I was a dial-up user for 8(14) years, cable user for 6 years, DSL user for 2 years, and FIOS customer for the last year and a half...

as for the 8(14).. well the 6 years on cable I had to keep a dial-up backup connection, so while i had cable.. i also was 'required' to have dial-up if i wanted a connection at 7pm in the afternoon..

was over 250 people on cable within a mile of me -- for faster-than-dialup.. we couldnt even get ISDN in the area for years... at 7pm when everyone was home and online, Ping times were over 1000ms, internet speeds were still 2mb+, but it took forever to get communications started.. and forget anything latency-dependant (voip, games)

Cable is often given both great and horrible reviews -- It's really dependent on your area, and how the infrastructure was set up there. A low populated area with a crappy infrastructure wouldnt be noticed, move that to an area with over 1000 people/sq mi. and you have 400+ cable connections trying to go down 1 pipe... but if they did their homework and set up the network ahead of time to support that many connections... you have a great cable network...

Nowadays the cable network is doing great.. almost everyone went to DSL/Fibre when it became available, most of us were tired of our cable provider :-)..

DSL, sure speeds are slower, and it is EVENTUALLY shared... when you're in a moderatly populated area, not having to fight for the connection is -great-... My history with connections:

dial up: ping ~350, ~8kb/sec transfers
Cable (early morning): ~150ms ping, 150kb/s+ transfers
Cable (afternoon): ~950-1200ms ping, 80kb/s transfers
DSL: ~150ms Ping, 3mb down 512k up
Fiber: ~50ms ping, 14mb+ down, 1mb+ up (i don't usually upload on the connection anymore)

Look at your neighborhood, ask around.. If your cable network is bad.. go DSL if you can. Make sure you talk to somewhat tech-smart people in some of your asking around.

Re:Time of day?

[Anonymous Coward]

When your phone line gets multiplexed into a T1's stream, your line get a guaranteed throughput of 64Kb/s OMG, DSL signals are put onto a T1, shared with other customers??? That stinks, because a T1 is limited to 1.536 Mbit/s... which perhaps is why DSL has such poor performance, particularly if you consider that multiple customers are sharing a relatively low-bandwidth T1.

Re:Time of day?

[alt-j]

I don't think that time of day applies. I ran some tests and found that they are actually bursing (or throttling, depending on your viewpoint) their bandwidth. Check out my pretty graphs: http://blog.alt-j.com/200704/comcast-bursting-uplo ad-bandwidth/ They're doing it for both uploading and downloading. It seems like they burst long enough to perform well on all of those online speed test sites.

Re:Time of day?

[g-san]

Dude, the 90s called, they want their 10Mbit hub back. Make sure to send back the coax, the T connectors and the terminator too.

Re:Time of day?

[Anonymous Coward]

And don't forget those ads they've got there. When I turn off flash in my browser, a lot of pages suddenly load very quickly (even on dial-up).

Re:Time of day?

[Anonymous Coward]

Of all the people I've seen arguing against net neutrality, I have yet to see one who knows what it is.

That includes you.

Re:Time of day?

[Short Circuit]

OMG, DSL signals are put onto a T1, shared with other customers??? If you'd bothered to interpret what I wrote, you would have noticed that I said voice data goes into the T1, and the DSL data goes on a separate network.

T1s were designed to carry over twenty-four or twenty-five (I don't recall which) simultaneous voice channels, at 64Kbps each. Incidentally, the same principle of multiplexing allows both the voice PRI and the internet service data to be multiplexed over a larger pipe.

I don't work at a telco, so I don't know if they have a T3 running out to the DSLAM, with the T1 and internet data multiplexed, or if they have the T1 running separate, with multiple channel-bonded T1s carrying the internet data. I imagine they've taken both approaches, depending on the area.

In any case, the internet data gets its own logical pipe outside of the voice T1 you latched on to.

Re:Time of day?

[evildarkdeathclicheo]

The whole "shared bandwidth" thing is a legacy misnomer when it comes to cable. Cable bandwidth is no more or less shared then DSL is these days. In the early days of cable internet service, the first CMTS systems were nothing more then giant bridges (lancity) and you did indeed share the bandwidth (and broadcast traffic) of everyone else on the network. This is where the whole public wives tale comes from. These days, systems are all docsis and you have dedicated QAM bandwidth to the CMTS device. After that, bandwidth is shared, just as it is on the rest of the internet, on whatever ingress/egress link your cable system has to their backbone, upstream providers, and peering points.

Re:Time of day?

[sumdumass]

This shouldn't be at all. And I think that is net neutrality doesn't find it's way in, some consumer protection laws will.

When an ISP advertises speeds upto 8 megs, they shouldn't be doing anything on purpose to artificially restrict those speeds. It is like me selling tickets to California from NewYork and boasting a travel time of 4 hours depending on the headwind to make it appear the fastest mode of transportation available and then purposely traveling slower to save fuel or whatever and getting you there in 19 hours instead. I neither delivered what I advertised nor was I truthful in my intent when advertising it.

This is what the FCC said they had a problem with when they didn't see much wrong with SBCs plans. They said the customer has to be able to get what they paid for. And if the advertisement says X megs speed, they better not do anything to make it slower then X.

Re:Time of day?

[Firehed]

Okay, I was talking about 22Mbps bi-directionally, and certainly not during spring break when nobody else was trying to use the connection, but 600KBps is cool too.

Re:Time of day?

[Firehed]

Except that this time, it was actually legal, legitimately free content.

Not that I'm implying that I've ever used Bit-torrent for anything else...

Re:Time of day?

[Harik]

Uh, yeah. I won't be the first nor last to point this out, but if Cable was truely "half duplex" then you'd have no upstream - period. Because if you'd get up from your computer desk, walk over to your television, and turn it on, guess what? Yeah, all those channels are broadcasting 24/7.

Thicknet/Thinnet lans used a single frequency for 'data transmission', modern cablemodems have vastly different upstream/downstream frequencies, and you are dynamically assigned a 'channel' thats much faster then your cap, shared with multiple other people.

Also, try Skype. Look ma, it works, full duplex. Guess you were wrong.

Re:Time of day?

[mokumegane]

Lol, my Aunt Carol gets to play with T1 connections every so often. I remember before cable and dsl... she put together these photos this guy wanted and attached them all to an email and sent them. It was like 50 pictures and sent to a dial-up connection... lmao! But really, Cox and Comcast suck in the south compared to the north. You were testing it in PA, which is north. Try it in Texas or something and you'll see what people are complaining about. :(

Re:Time of day?

[volkris]

Of all the people I've seen arguing FOR network neutrality I've seen about twenty definitions of what it is. The pro people don't know what they're talking about either.

They only agree that the evil phone companies are after poor innocent Google's money.

Re:Time of day?

[raju1kabir]

DSL: ~150ms Ping, 3mb down 512k up Fiber: ~50ms ping, 14mb+ down, 1mb+ up

Your DSL line had 100ms more latency than FIOS? That's just weird. DSL adds about 8-30ms depending on the type of link.

Re:Time of day?

[cablenetguy]

That's a fallicy. A cable modem is shared at the HFC Node. DSL is shared at the CO. All of your bandwidth is shared at some point, even if you pay for a dedicated fiber connection. There are newer solutions for cable to expand the HFC bandwidth by adding more upstreams with split-nodes and even more possibilities with Docsis 2.0 when it is released.
If you're having slowdowns at some particular times your cable company needs to look into their bandwidth utilization.

Giving Comcast Props

[tenchiken]

It kills me to say something nice about the brood of bloodsuckers that are Comcast, but I can verify that Pockets of Comcast's net are seeing huge increases. In particular, I have seen speeds of 19-22Mb/s burst to testing sites, and almost 2.0MB/s non-bursting.

That's in the Denver region using both speedtest.net and DSL tools.

Give credit where it's due, but Comcast does appear to be amping up the bandwidth hugely.

Between this and the Zimbra announcement, Comcast has firmly passed Qwest as next to last evil corporation.

Re:Giving Comcast Props

[Anonymous Coward]

As an unfortunate subscriber of Comcast, I'd love for their bandwidth boosts to be true. However, I just took this test today with results of 16784kbp down and 1728kbps up. It'd be wonderful if this were true, however the fastest downstream I've ever achieved is about 1100 KB/s (~half of what I got on that test), and I only achieve that in extremely rare circumstances. Just a handful of times since I've used them these past few years. My upstream? No, I don't even think I've ever reached half of that.

Not to mention Comcast's unbelievable unreliability, poor customer support, and general poor service as an ISP. I lag more than any of my friends in online games, on both computers I own, and my connection has a tendency to quickly reset a couple times a day, kicking me out of a game. I sometimes lose signal for a couple hours at a time, though this occurs every couple of months, thankfully.

I don't recommend Comcast to anyone.

Re:Giving Comcast Props

[Icarus1919]

To everyone patting Comcast on the back for a few scattered examples of decent, above board service, I advise you to check The Consumerist for a bit more information on Comcast's practices.

Re:Giving Comcast Props

[Novanix]

Actually I get really high burst speeds, but if you a long download of >10-15 seconds it will drop to your 6-8 megabits a sec rate. It confuses all the speed testers as basically any large transfer for the first so many seconds will go very fast, and its just after that the speed really drops. Best way is to just downlod a large file and see what the speeds are like 20 seconds in.

Re:Giving Comcast Props

[ajf442]

I have been using Comcast for almost 5 years now...it was Comcast until 6 months ago...it is now Time Warner (road runner). I ran the speakeasy test on the Dallas Server and got 5.6 Mbs download and almost 500kps download. Even though Time Warner bought them out here...it is basically the same service.

Re:Giving Comcast Props

[MBCook]

I have Comcrud. They bought out my local cable company years ago and raised rates, trashed service, capped the internet speeds, and all sorts of other fun stuff. I have something like a 6 mbps connection if you listen to them. I've seen it up as high as a little over 1 mbps (from speed tests and such). As it is now, every speed test I can find is maxing out at about 300 kbps down and 350 kbps up, which is what I was getting hours ago when this story first hit the firehose.

Jerks.

I love monopolies. My only other choice is unreliable ISDN at 3x the price, or a (partial) T1 which costs even more.

Re:Giving Comcast Props

[iamdrscience]

I can verify that Pockets of Comcast's net are seeing huge increases. In particular, I have seen speeds of 19-22Mb/s burst to testing sites, and almost 2.0MB/s non-bursting.

I've had Comcast and I found their download speeds to be quite usable and the occasionally help of burst speeds like you mentioned is nice as well. However, their home service is still only available with 40kB/s (320kbps) upload bandwidth which is shit. And to make things worse, it seems that the more of that upload bandwidth you use, the less download bandwidth they make available to you. When using the full upload bandwidth the connection becomes almost useless for downloading.

Re:Giving Comcast Props

[Tuoqui]

I'm a bit skeptical on this...

It could just be that they are using QoS and Packet Shaping to speed up connections to these 'test sites'. Allowing for 'bursty' traffic to break formal limits for 5, 10 or even 60 seconds a test takes. This makes people happy going 'Hey cool I'm getting X-speed here when I'm only paying for Y-speed'. A customer placebo if you will.

That is the problem with net neutrality not being respected by ISPs, they can engage in these sorts of behaviors without anyone knowing if they're getting legitimate results or if their packets to X-test site are being prioritized over anything else.

Re:Giving Comcast Props

[Erpava]

Did you even ask for the 6mbps connection, or did they just stick you with it? When Comcast bought out Time Warner in my region, they gave me the "premier" package without even asking if I wanted it. $10 more per month for the same terrible upload speeds? No thanks.

Re:Giving Comcast Props

[Eric in SF]

Two things:

I have Comcast (8/768) in Cole Valley, San Francisco, and I have also noticed the speed increase. Uploads to my website are now cruising at 140kb/sec, occasionally dropping to 90kb/s. No complaints here! I performed a dslreports speed test recently and it also reported some Korean or Scandanavian-class bandwith numbers - the highest I've certainly ever seen in my time with broadband.

Second - it's my understanding that as you saturate the uplink connection (max out uploading a file) on a consumer-grade connection/router, you interrupt the normal control-channel "Chatter" of web browsing. Basically, the "I got it" packets are stuck due to the saturated uplink, and you don't get the next packet until the acknowledgement makes it.

I could be completely wrong - I am by no means a networking expert, so if this is wrong, be gentle.

Re:Giving Comcast Props

[jlarocco]

Give credit where it's due, but Comcast does appear to be amping up the bandwidth hugely.

I got a massive bandwidth increase for about a month right after Comcast took over Adelphia and started blocking bit-torrent. So don't get too excited, they're just using the increased bandwidth to hide the fact they're breaking your service somewhere else.

Re:Giving Comcast Props

[Anonymous Coward]

When using the full upload bandwidth the connection becomes almost useless for downloading.

You're an idiot.

Re:Giving Comcast Props

[tenchiken]

Which is why I documented non-burst speed above.

Re:Giving Comcast Props

[MBCook]

We asked. We had a little local outfit that was great. Then Comcast bougt them out. We've changed our plan once or twice (like adding HD when I got an HDTV, and lowering the cable speed since we were no where near that speed). Things have gone quite downhill, but they didn't do that do us. As I said in my above post, we have service with them because they are the only game in town for us.

Re:Giving Comcast Props

[Blkdeath]

Second - it's my understanding that as you saturate the uplink connection (max out uploading a file) on a consumer-grade connection/router, you interrupt the normal control-channel "Chatter" of web browsing. Basically, the "I got it" packets are stuck due to the saturated uplink, and you don't get the next packet until the acknowledgement makes it.

Actually you're absolutely right. Any TCP (connection based) protocol will suffer from a saturated uplink to the point where it can become unusable depending on how greedy that client is about your outbound packet flow for exactly that reason; every time you request a packet you have to ACKnowledge its receipt. If enough packets go unacknowledged the sender will stop transmitting to give you a chance to catch up. At the point where you have your outbound 100% utilized, your connection becomes useless for even the simplest of tasks.

The reason residential connections tend to be asymmetrical is the fact that ack packets are much smaller than received data and the nature of the Internet nowadays (even with our media rich content) is that people are downloading files in the 100s of megabytes (a 720p movie trailer on Apple's website will generally run about 120 meg) but only need to send 1/10th or less back to the sender to acknowledge receipt.

Some time ago my room mate got heavily into filesharing networks. Upon trying out a new client he found he was getting better speeds. Well, that's because his download speed was based on a ratio to his upload speed and the limits I'd set in his old client naturally were not set in the new one.

I tried to connect to my home server from work to check my e-mail and found that the connection was slower than a 14.4k dialup modem. I ssh'd in, found the source of the traffic and promptly shut down all direct-to-Internet traffic from his machine and poof! No more link saturation!

Re:Giving Comcast Props

[Firehed]

As a slashdotter, I doubt it, but could you have confused Mbps with MBps?

When I was on Comcast (actually, come to think of it, I think I am here too), I usually got speeds right around what was advertised. Hell, I got a speed boost when they took over Adelphia.

Re:Giving Comcast Props

[vought]

In particular, I have seen speeds of 19-22Mb/s burst to testing sites, and almost 2.0MB/s non-bursting.

I have seen fairly consistent rates in Silicon Valley (San Jose) of 14-19Mbps on Comcast's basic Internet service.

Only 1.5Mbps up.

Re:Giving Comcast Props

[AaronW]

I just ran the test and got 21,499Kbps down and 601Kbps up, at 7:45pm from Fremont to San Francisco. Granted, I pay an extra $10/month for more bandwidth (for more upstream bandwidth) but lately I've noticed a nice improvement in downstream bandwidth, given that it's supposed to be 8Mbps. I'm not complaining.

Re:Giving Comcast Props

[AaronW]

I have had few problems with Comcast. It is not unusual to get over 1.5 megabytes/second when downloading large files from a fast site. It's expensive, but much faster than the competition around here.

The speed test reported 21,499Kbps down and 601Kbps up.

Re:Giving Comcast Props

[dabraun]

I have to give Comcast props as well. I haven't used speed test sites, but I've observed my download speed for actual files I want, go from a few hundred KB/sec to over 1MB/sec in the last few years. Ok, I realize that this isn't as dramatic as what some are reporting, but I'm getting what I believe is about 12mbit/sec when I originally signed up for something more like 3-4mbit/sec - and all the while the price has not changed. Sure, it's a little slower in prime time of the evening, but it's still quite fast, certainly a lot better than I ever got with my Verizon DSL.

Re:Giving Comcast Props

[pen]

The bursts of traffic are not designed for the test sites. Most people don't care about the test sites. They're designed for downloading that 5MB Word document or PDF that someone e-mailed you or linked you to.

Re:Giving Comcast Props

[whoever57]

I have also seen good speeds (>15Mbps) when downloading using Comcast. However, upload speeds can be problematic. Just today, I was uploading using sftp and it started out at 150kBps and quickly dropped to ~40kBps (somewhat greater than 320kbps, depending on the overhead)

I have also seen slow ssh speeds to my work LAN, yet when I go via an external server, the ssh speed is much better.

Re:Giving Comcast Props

[Alan Doherty]

>>When using the full upload bandwidth the connection becomes almost useless for downloading.

>You're an idiot.

actually the explanation above is generally correct {but not observably so on a well setup network}

as the effects of uplink saturation should be mitigated by intelligent queueing on the router
{this is not always the case}
        small packet & tcp sequence acknowledgments etc should get priority on a saturated link and icmp-slowdown-transmission-rate packets should be sent to the client pc causing uplink saturation,

but with badly configured routers/firewalls & badly configured "personal" firewalls and p2p clients on client pc's these messages often do not get through, or are not watched-for/implemented by the client application

Re:Giving Comcast Props

[symbolic]

I have Comcast, and I also had a BAD resetting problem - I'd get reset on almost every game I tried to play. Sometimes I'd make it through a couple of hours, but those were usually the exceptions. In looking at all possible causes, I upgraded my router's firmware, and all my problems disappeared. It appears as though it was the damned Linksys router causing the problems.

Re:Giving Comcast Props

[timeOday]

20 mbit burst a big bonus for apps like google maps. At a hotel recently I noticed maps was awfully sluggish, but checked and realized I was pulling half a megabit, which used to be quite a bit of bandwidth! Even cnn's homepage is most of a megabyte. Man, the web has really gotten bloated. Any dial-uppers out there? How do you do it?

Re:Giving Comcast Props

[Short Circuit]

Oddly enough, dial-up ISPs seem to go further. I ran a speed test on my dial-up connection because I wanted to see how much bandwidth I had left over while an automated Windows Update download (Which I had to put up with for five days...) sucked up most of my downstream. While I only got 6Kb/s down, the speed test reported 150 Kb/s up.

Yeah, try telling me a 56K modem, connected at 28.8 Kb/s, is transmitting data at 150 Kb/s.

Re:Giving Comcast Props

[dreddnott]

I've been using dial-up at home for all of ten years. They won't even pave our frontage road let alone string cable or DSL out here.

How do I do it? Well, I suppose I'm just used to it. I'd say I use Opera to cope, as it's such a snappy, efficient browser, but I've been using Opera for at least eight years (version 3.something), since before the Internet was bloated, so I guess that doesn't count. I do set the Opera caches to their maximum sizes to minimize the horror of redownloading the static content of websites I visit frequently. Ah yes, I remember now. I really started using Opera because the download size was (and still is) extremely small compared to the other browsers.

I never go to Youtube, don't use MySpace, Google Maps, or other silly websites that break if you don't download the content fast enough. I browse Slashdot with more minimalist settings, with my threshold set at -1 and -all- posts displayed inline so I don't have to click on any comments to actually read them.

On other web forums I usually use the simplest scheme (Relicnews forums has a great setting called "Adminimal") and if I have trouble on a particular site or I get impatient (usually after half an hour of waiting) I'll turn image downloading off and manually download the images that I think I'll need to see.

Google displays 100 search results per page, and I set all forums to the maximum number of posts per page. Display Print Version is a welcome feature on those retarded hardware review websites. It's much better if I download it all in one go, there's a lot of overhead for me in CSS/HTML these days.

On the whole it's not so bad, over the last decade I've still managed to acquire a substantial collection of MP3s (I was in fact one of the 300,000 names on Metallica's hitlist for Napster). Dial-up first started to frustrate me in 2001 when I started making anime music videos. I used my own DVDs as source material but uploading the videos was quite a pain. I remember the first time I downloaded a large file on dial-up, I think it was around 32 megabytes. I was very proud of myself.

I used to play Unreal Tournament over the Internet when the game was very popular (1999-2002). I think the best ping I ever got was around 150-160, and I was so used to playing at over 250ms that I won every round on that server. I was very competitive at LAN parties (ok, I won grand prize at a tournament for budding game developers) but rarely got the opportunity to play at one. I think I could reclaim the glorious days of my youth if I had a better Internet connection...or if this podunk wasteland they call the High Desert of Southern California actually got some good local events going on.

Re:Giving Comcast Props

[anagama]

I'm on Comcast because its my only choice. Seattle is 90 miles away. Speakasy results to and from Seattle: 2768, 362. New York city is farther: 2229, 360. Basically, my network sucks at any distance. Worse, because I don't buy cable TV, I have to pay an extra $5/month for my service. After all is said and done, that's $60/month for internet alone. I long for the days when Comcast has some competition here.

Re:Giving Comcast Props

[dmsuperman]

I wish I could get that. Burst speeds I get max 3mbps, with an advertised 6. Average speed is below 2. Comcast blows as far as speed goes, but they do have excellent customer service. They just switched to Time Warner in our area, and Time Warner BLOWS all together. Worse connection, and the customer support is horrible. Couldn't check my email because it was too long (it was only accepted because it came with the change to Time Warner) and it took 7 calls, with 2 level 3 priority support tickets and to talk to a regional supervisor to get it remedied. Just wanted to check my email :s

Re:Giving Comcast Props

[u-235-sentinel]

Give credit where it's due, but Comcast does appear to be amping up the bandwidth hugely.

True however you can only get into trouble with their Security / Assurance Department faster now. They have been terminating people's accounts without giving them adequate notice. Also they are unwilling to define what "Acceptable Use" is. Basically it's a crap shoot. You don't know if you are violating AUP/TOS and they won't tell you how much bandwidth is acceptable. Only will give you examples of what isn't acceptable.

It's like the Government telling you driving 100 MPH is unacceptable. Well that's fine and dandy. I think most people will agree with that. But not providing guidance in the DMV manual of the speed limits where no posted limit is specified, now that's darn right wrong.

I would strongly urge caution. You simply don't know if you are going to be next by downloading at those speeds.

Re:Giving Comcast Props

[mdm-adph]

I second this -- I live in Bumfuck, Alabama (middle of nowhere, seriously), and I'm getting 8Mbp/s down and 1.7Mbp/s up with Comcast. Something's fishy.

Re:Giving Comcast Props

[Valdrax]

If you don't mind, could you tell me how much was unacceptable and what general area you live in?
The *only* thing keeping me from switching to cable is fears that my BitTorrent usage will get me slapped.

I mean, the *only* thing. I have a list of reasons to boot AT&T as long as my arm, and I'm still stuck with them.

Re:Giving Comcast Props

[angus_rg]

The other problem you have is those tests may show latency/lack of bandwith, they just don't show where, and there are a lot of hops between your ISP and your detination server. If you run a traceroute simultaneously and see latency near the backbone, you might be able to blame your ISP. Even that may not be accurate since it is only a snap shot.

Re:Giving Comcast Props

[Khaed]

I tried to connect to my home server from work to check my e-mail and found that the connection was slower than a 14.4k dialup modem. I ssh'd in, found the source of the traffic and promptly shut down all direct-to-Internet traffic from his machine and poof! No more link saturation!

One day, seemingly at random, sites would refuse to load, the internet was slow and sluggish. I thought it was just my provider sucking ass, because that occasionally happens. When, the next day, it didn't stop, I asked if anyone else plugged into the home network was downloading. The answers were all no. Another day passed, and I noticed that it only sucked ass when my sister's boyfriend had his laptop plugged in to the router. So when he leaves, he leaves the laptop plugged in and on. Why would he do this, I wonder? So I check his laptop, which isn't in any way secured from me accessing it physically.

He's got bittorrent going on, and he has no restriction on his upload speeds, and they are literally maxing out all of our available upload. I check the numbers, divide by three (the number of computers on the network), and set his maximum upload to half that through the client. Did the same for download: took the max, divided by three, gave him half. I never said anything to him, so for all I know, even now he's got a crappy upload speed. It's been over a year since I did this, since then, I've changed providers and more than doubled my available bandwidth. But the way I see it, he point blank lied to me about downloading using *my* connection and *my* net sucked until I figured out what it was. So I hope he's enjoying his 50kb/s maximum download in bittorrent.

Had I thought of it, I should have changed his screensaver to "marquee" and had it say "At least put up a screensaver password, numb nuts!" ...And passworded it.

Re:Giving Comcast Props

[Thundersnatch]

every time you request a packet you have to ACKnowledge its receipt

TCP has supported Selective Acknowledgement for a very long time now. It's even on by default in Windows (since 2000, I believe). Essentially, this TCP option allows the receiver to ACK a range of packets received with a single ACK packet. It even allows multiple ranges to be ACKed, so the sender knows exactly which packets to retransmit. I would assume Linunx and BSD support it as well.

Re:Giving Comcast Props

[Blkdeath]

every time you request a packet you have to ACKnowledge its receipt

TCP has supported Selective Acknowledgement for a very long time now. It's even on by default in Windows (since 2000, I believe). Essentially, this TCP option allows the receiver to ACK a range of packets received with a single ACK packet. It even allows multiple ranges to be ACKed, so the sender knows exactly which packets to retransmit. I would assume Linunx and BSD support it as well.

You'll note that I never said every packet has to be individually acknowledged, only that every packet has to be acknowledged. The problem with torrent networks is that they provide so much backscatter and generate so many packets they quickly outstrip home user grade Internet connections.

But thanks for the reference. I'm sure some readers found it useful. ;)

Re:Giving Comcast Props

[IAmRenegadeX]

Haley's Hot House is still a UT hotspot. The game still makes it a bit easier on high-ping/modem users. Drop by, sometime!

ISPs up to dirty tricks

[sc0ob5]

Who would have thought?

Boost & caching

[masterz]

It's possible they cache the data. In most cases, the test will say the results are unreliable. Comcast also offers a service where they don't artificially cap short bursts of information. I assume that's "PowerBoost"

Re:Boost & caching

[Repossessed]

Powerboost is probably what's causing this, yes. I *hate* that thing. Messes up my benchmarks when I'm testing ftp servers. Does not, as far as I can tell, do anything that's actually useful. I also know one person whose service is grandfathered in from a company (Adelphia) that Comcast bought out. He gets about 15-20 Mb/s depending on the time of day. As far as I can tell, that's not power boost related (at least I get a legit speed off the same test site).

me too

[Anonymous Coward]

Anonymous for obvious reasons...

I see this from Comcast too, I attribute it to the cable modem that I purchased elsewhere. I think Comcast doesn't know how to throttle it.

Good Speeds

[GizmoToy]

Those are some pretty damn good speeds. I was lucky to get 4Mbps at Time Warner when their max was 6Mbps. I switched to a 5Mbps DSL connection which ended up being faster. I wonder if there's much of a difference now that they bumped it up to 8Mbps. And it seems with Comcast apparently bumping up to 16Mbps according to another poster, a connection speed race could be on. Maybe it's time to look into switching back?

simple

[ILuvRamen]

their "power boost on large files" reeeeeally sounds like a compression feature like a very few dialup companies use. Basically you could download a 1 GB file of the letter A over and over in a few seconds on dialup with that compression. Don't most speed tests just download a giant gob of repetetive data? Maybe that's why. Just a theory though, dunno if that's what's up. I do know for certain that Comcast sucks.

Re:simple

[Anonymous Coward]

Well, no. Typically the information being transmitted is uncompressible or only trivially compressible.

Well thank god we still have TW in Raleigh

[gelfling]

What with their rock solid 3.7MB down and 374KB up - Anything more would be Goddamn communism.

Re:Well thank god we still have TW in Raleigh

[kaufmanmoore]

TWC, Bellsouth and Clearwire.....NC is still in the dark ages, not 1 offering of double-digits down or even 1mbit up.

Re:Well thank god we still have TW in Raleigh

[Anonymous Coward]

I'm in the Greenville, NC area. With my dsl though embarq I get a measily 5.06MBps downside and 778KBps up solid downside can easily sustain 6MBps on many downloads. Also get 99%+ on qos line test. That on my 5down/778up plan. Think my uploads are pretty good when I compared them to many other offering in my area.

probably not fibbing

[SocietyoftheFist]

I had 6Mbps service through them several years ago and routinely downloaded ISOs at 10Mbps.

DOCSIS 2.0 Plus

[Fallen Kell]

It all depends on the cable modem that you have. Some of the new motorola modems, like the SB5120, do not have the ability for Comcast to limit as much as they sometimes like. Comcast themselves has not been too worried about it as long as the network segment you are a member of is not over-crowded. They see it more like a new benefit which allows better competition against FIOS. Personally I average 25-28mbps on my modem.

Here is a current snapshot:

Re:DOCSIS 2.0 Plus

[Fallen Kell]

DOL... forgot the link is a pic/link... well, it was 20,000+ down, 1,500+ up...

Re:DOCSIS 2.0 Plus

[Daikak]

what i have noticed personally is now that i'm using comcast digital voice my speed tests do this also. kinda wondering if comcast even knows how to handle their new modems too then heh.

Re:DOCSIS 2.0 Plus

[hansamurai]

So are you using the SB5120 then? Is it worth the 50 bucks at Amazon or Newegg? I torrent a lot with moderate internet surfing.

Re:DOCSIS 2.0 Plus

[JimDaGeek]

Why would the cable modem make a difference? Wouldn't the ISP's be doing the throttling based on your MAC address?

I recently upgraded my 5 Mbps TW cable connection to 8 Mbps for an extra $10/month. Within 1 minute or so, I had the faster speed. When I called they just asked me for the MAC address on my cable modem.

Re:DOCSIS 2.0 Plus

[Anonymous Coward]

At least some throttling must be done on the modem end, because otherwise there would be nothing preventing you from flooding your local segment. They can drop the traffic after it passes the local segment and this would generally encourage the protocols to send less data, but there's nothing saying your protocols must play nice. Of course everything that uses TCP will, but not everything does.

Once they have modem-end throttling in place, it's easy for them to be a little lazy and decide that's all they want to do.

Re:DOCSIS 2.0 Plus

[Fallen Kell]

Yes, I have that modem myself. I went from an older model that Comcast upgraded me to when they upgraded to 6mbps in the area. It was an older motorola surfboard modem. A friend at work read about the SB5120 and upgraded to it and showed me the results, which were dramatic for him. He was seeing 30mbps at times. I then purchased one for myself. I did speed tests before and after the change. I went from seeing approx 8mbps down and 320kbps up to 28mbps down and 1.5mbps up. This was over the difference of a 10 minute call to Comcast to get the new modem activated... I still normally see in the 20's all the time, with peaks in the low 30's. Best $50 I ever spent.

Re:DOCSIS 2.0 Plus

[heelios]

What happens is that the ISP pushes a config file to your cable modem. The throttling is done on your end which is why is it possible (although quite unwise) to illegally uncap the connection altogether.

Re:DOCSIS 2.0 Plus

[jaden]

Hi, I don't think the cable modem should make a huge difference as the cable companies push a config (profile ?) to each modem setting the max upstream / downstream limits. When on cable I was never fortunate enough to be getting higher up/down speeds than paid for... but when I was being capped below my service plan I used this great tool: DocsisDiag to grab the config settings currently on the modem. Good info to have when arguing with tech support. Not sure how the profile would look for Comcast with speedboost... but it'd be interesting to that or the config from your modem with 25-28mbps. -J

Re:DOCSIS 2.0 Plus

[Anonymous Coward]

What about the router attached to your modem?

I bought my router in 2002. The WAN is 10Mbps, while the LAN is 100Mbps. Currently my ISP, InsightBB, has us on a 10Mbps/down 1Mbps/up listing. (What they claim). I have not looked into a new router as I figured I might not see much of an improvement. The theoretical max is about 8Mbps, due to network congestion, correct?

I am assuming that if I upgrade my router to one with a WAN of 100Mbps, I might see my speeds jump closer to 10Mbps, and maybe burst higher. Am I correct on that?

For the record, I test out at approx 6Mbps/down and 950 Kbps/up using a test point 200 miles away, and I range 5.5-6.0Mbps/down and 890-950Kbps/up with test points 500-1000 miles away. I have gotten higher down speeds at other times, up around the 7Mbps range in previous months.

Re:DOCSIS 2.0 Plus

[Anonymous Coward]

Actually I have a old SB4200 that is running on Comcast's DOCSIS network, and I can get download speeds of between 20 - 30 Mbps. That's because the config files that Comcast is pushing out to modems now allow bursting. You can actually find information online about how to look at what config files Comcast is pushing out, and how to read them.

So the modem should not make a difference. When you turn on your cable modem, it tries to register on the cable network and then is told which config file to download based on its MAC address. This is the same for ALL cable modems that run on a DOCSIS network.

Well...

[Creepy Crawler]

Try a real-life saturation test: Bit Torrent.

Tell whatever client you're using to go all out.

If they block it, use Jidgo to download ALL of Debian (but that only tests download, not upload).

Re:Well...

[JustNiz]

No its no good using p2p. Most if not all the ISP's do traffic shaping (read: bandwidth throttling) on p2p traffic.

Re:Well...

[Anonymous Coward]

Sacramento Area Comcast Real world bandwidth here tops at about 1.6MB/s down and about 95K/s up.
I have had same ip for ~6 months, and have had maybe 5 mins of downtime in 2 years.
Services is advertised at 8mbps(what i pay for), i get ~ 20mbps on speed tests.

I agree

[certain death]

with the adage of "Give credit where credit is due"...Now, I had Comcast from 2004-2006 and if they could keep my freaking line up, the speed was about 6 to 7 megs consistently. Again, I say, IF they could keep the line up. Speed of broadband is only half of the equation. The second part is service. If your customers can not even connect due to shoddy or shitty, you pick, equipment, and cable infrastructure, then does the speed even matter? I now gladly pay $35.00 per month for DSL at 6meg, and I have not had an outage in over a year.

Switch-over in Houston, TX

[DigiShaman]

Those of us living in Houston, TX should keep a sharp eye on Comcast. We used to have Time Warner, but this month the entire network has been switched over to Comcast. The before and after service analysis will be interesting.

Again, keep a sharp eye open. We can't miss this opportunity.

Re:Switch-over in Houston, TX

[cepher]

As another resident Houstonian, I haven't been too pleased with Comcast's service so far. We have had at least 2 outages since the switch (July 29th) in Clear Lake, and Comcast's support lines were unreachable and their web service had 119 people in line ahead of me. Plus, the lines all over Houston are the same. I have yet to see any people doing upgrades to the lines or seen any changes in speed; except lack of service.

Re:Switch-over in Houston, TX

[DigiShaman]

A few of my customers located in Pasadena, TX also had an outtage on that day (June 29). I was able to contact tech support through the business class support (as that was they account they have). Turns out their LDAP servers bombed out from the switch over. We've also experienced some strange DNS issues too. I suppose it was to be expected honestly. A switch over isn't always going to be flawless. Good news however, our customers and myself haven't noticed any more issues after that day.

PowerBoost uses compression

[rdean400]

I'm wondering if the speed test uses a data block that is more compressible.

Re:PowerBoost uses compression

[Anonymous Coward]

Can you back that statement up? I've seen no evidence that PowerBoost uses any sort of compression (it works fine on files that are already compressed) and if they had some magical compression technology then they ought to just leave it on all the time and either get increased speed for free or less network utilization.

PowerBoost

[the_cowgod]

The Speakeasy test regularly reports over 20Mbps on my Comcast connection. The "PowerBoost" feature allows basically uncapped speeds for the first ~15MB of a transfer, then it drops down to the normal 6Mbps. I can easily see this effect when doing large downloads with my UsenetServer account. It does inflate speed test results, but Comcast does not appear to be favoring the test sites in any way.

So that's where my missing bandwith is!

[letsgolightning]

Since many of us are lucky to get 75% of what we pay for (during off-peak times), it's good to hear of someone getting more than they paid for. My completely random guess is that someone may have misplaced a decimal point while capping lines in the area. Perhaps the guy in charge lives next door and decided to give himself some free speed?

Faster Speeds

[steeb]

If you have TV, Voice and Internet, you will get faster speeds. If you subscribe to the voice and internet and claim when you surf the net (you tube, myspace) your phone breaks up.. They will up your speed no problem. Buisness Lines - I've seen people get 10mb down for having a buisness account with comcast. And then it all depends on your area. If you call and complain at night that you're getting less than promised speeds, they can up you. my 2cents

Re:Faster Speeds

[Anonymous Coward]

In Northern California, towards end of last week, running a test, got a download speed of 184 kbps on COMCAST. It is so bad that I cannot use my VOIP phone service, even when 'net is not being used. Weird as it may seem, there seem to be some correlation between outside temperature and how bad the 'net speed is (it has been unseasonably warm here)

Re:Faster Speeds

[tylernt]

Weird as it may seem, there seem to be some correlation between outside temperature and how bad the 'net speed is (it has been unseasonably warm here)

Probably a cabling issue on your end. Our cable service would suck every time it rained, until we had them come out and bury a new cable from the box to our house.

Re:Faster Speeds

[pyrrhonist]

If you call and complain at night that you're getting less than promised speeds, they can up you.

Steeb upped his speed, so up yours!

Re:Faster Speeds

[Anonymous Coward]

> there seem to be some correlation between outside temperature and how bad the 'net speed is

That's not uncommon. I live in SC, and it gets very hot and humid here. A lot of the cable amps get flakey when they overheat. I've seen plenty of cable connections that worked fine at 7AM when you watched the morning news but the 6PM evening news was unwatchable. What you're seeing is normal for the garbage type of equipment that Comcast buys. I work for a company that makes cable equipment that works well for years at 120 degrees so I find their decision to intentionally buy equipment that doesn't work very annoying. Locally I have Charter, and they buy the same crappy equipment that Comcast does. They see no problem with saving $500 on an amp upfront to only have to spend much more than that making dozens of service calls to angry customers. The management at those two companies is all about the shortterm gain to ripoff the stock market with no thought to the long-run.

The other problem that probably doesn't affect you as much in CA is moisture. Around here with the hard rains and humidity, the cable goes out several times a day. The cable companies are too cheap to use NEMA boxes or watertight connections.

Re:Faster Speeds

[Anonymous Coward]

Good point about the moisture. That's why here in FL you can't use cable if you have to depend on it to work after the typical afternoon thunderstorm. Cable just plain doesn't work when it is raining or for hours afterwards. DishTV is great. It works in all but the heaviest storms, and for Internet access while DSL isn't perfect, it at least doesn't quit everytime it rains. If it does quit, just call Hell$outh to complain about your POTS line and they will fix the problem. Hell$outh has to by law. The cable company will just laugh at you if you complain about poor quality. As they enjoy reminding you, if the cable works at least one minute for seven days in the month then they can bill you for the entire month. I think sometimes they're just playing games to see how bad they can make the service and still get people to give them money.

Faking performance?

[lancejjj]

In contrast I have heard bad things about Comcast. So now that I am up in the Philadelphia PA area visiting my parents, I decided to test out the speed and reliability using the Speakeasy speed test. I've seen the same thing in the other Comcast markets (but not all over). What would it take for Comcast to throttle all traffic except that going through Speakeasy's "speedtest" IPs? I'm imagine very little, and it likely could make some people feel good about Comcast's performance. That's not to suggest that Comcast is actually faking it - I'm just saying that it could be done. Clearly, Comcast wouldn't do this - it'd be a stupid risk for them.

In any case, maximum download throughput is only a part of service quality. You also need to measure network latency, upload throughput, DNS quality and performance, DHCP reliability, mail services, etc etc.

Network download bitrate is like yesteryear's "MHz wars" - a nearly meaningless value unless the whole system is understood, but a fabulous marketing device.

FYI - my bill for Comcast Internet service is over $55 per month. More than I pay per month in electricity.

Re:Faking performance?

[Anonymous Coward]

It's not giving you more bandwidth for the tests. It's just the tests are so short. When I initially start a usenet download it goes up to 12mbps. After like 10 seconds, it's at 6mbps exactly which is the amount I pay for.

Re:Faking performance?

[musicmaker]

If you only pay $55/mo in electric you must be living in either a very small place, or somewhere where electricity is unbelievably cheap.

Re:Faking performance?

[wtfpgh]

I guess electricity is unbelievably cheap here. I paid $43 last month for my electric bill, which (same as the GP) is less than my Comcast bill.

Re:Faking performance?

[lancejjj]

$0.15 a kilowatt-hour. The trick is to turn off stuff when you're not using it. Having the A/C run all day when no one is home is just stupid.

I swear to God

[Travoltus]

Comcast had boosted us up to 28mbps download speeds according to Speedtest.net. I routinely confirmed it with download speeds of around 2.9 megabytes per second of total incoming download bandwidth during quieter times of the day (midday, night time). I should have been getting 3.5 megs per second, but I attribute the difference to a shiteload of 24/7 Youtube-like traffic.

They said they were doing some kind of trial in my area, and that it would end.

Well, it did end, and I went back down to a max performance of 18mbps on speedtest and a real world cap of 2 megs per second.

I was crushed. So crushed. Whatever was I going to do with 2 megabytes of download speed per second and 100 K/sec of upload speed? LOL.

Oh and I'd had a total of 45 minutes of internet downtime since I signed up for them in 2004.

I'd have kissed Comcast's Comcastic feet if I didn't know they were a corporation and that corporations are given to turn on their customers like a rabid dog at a moment's notice. Especially one like Comcast which has no competitors anywhere near their class (SBC DSL tops out at 3mbps in this area and God knows what DirectTV can provide for internet service).

Thank you, Comcast, for all the great service you've provided me since the day I signed up... but don't think I didn't keep a suspicious eye on my bills, my net performance, and your policy enforcement.

I'd go back to Comcast given my experience, but I've actually graduated to a DS-class line. Good for business, good for writeoffs.

Download a linux distro

[Raleel]

Someone above mentioned using jigdo to get all of debian, but even using wget to get a full dvd or cd set of some distro will provide you with a good data point. That's been my standard test for quite a few years now.

Re:Download a linux distro

[grcumb]

Someone above mentioned using jigdo to get all of debian, but even using wget to get a full dvd or cd set of some distro will provide you with a good data point.

I beg to differ. Downloading a single file is only indicative of how fast a particular connection is between two particular points. P2P, if it's allowed to, saturates the network with two-way traffic to numerous end points. If - and this is a big if - there were no constraints on P2P traffic at large on our networks, I would consider it a near ideal measure of TCP traffic capacity in the real world.

In fairness to your comment, you characterised that single download as a 'data point'. Strictly speaking, that's a valid statement. How to get enough of these data points to provide useful insight isn't adequately answered by either of our suggestions, and frankly I'm a little doubtful about the online speed tests being discussed here, too.

Re:Download a linux distro

[ydrol]

even using wget

aria2 -s 5 big.arse.file

that made me think how few UK "big ass" porn titles there are. It's right but it just doesn't look right.

Comcast Fluke?

[micksam7]

I've noticed random increases in speeds on Comcast, expecially during weekend night hours.

It's totally random, but occationally the 5 Mb/s throttle here totally disappears and I get 20/30 Mb/s and more continously, then just as suddenly it's back down to 5 Mb/s a hour or so later.

I suspect it's just a fluke in their hardware, or they could be testing links and temporarily increasing the cap everywhere. Whatever it is, it never lasts long, maybe a day at the most before everything goes back down to speed.

It's nice knowing they have the capability to go up that far, for when FiOS gets big in the area and they increase limits to keep up.

Comcast in Washington DC

[Southpaw018]

I have Comcast in the DC Metro area (Maryland side). Just ran a test, and I got 5856 down and 363 up to the DC server at Speedtest. My connection is advertised as 6Mbit - seems to be spot on.

Re:Comcast in Washington DC

[Anonymous Coward]

Hi. I have Comcast :::spit::: here in Atlanta. Basic cable TV with the box, and Internet service.

I just tested and got 18383 down and 1168 up.

I am conflicted about this.

Re:Comcast in Washington DC

[maotx]

I do as well (Greenbelt.)
Speedtest.net tested against Frederick, MD ~13Mb down.
Sites just a little further away, such as Ohio, give me ~5Mb down.

Too bad Comcast doesn't do much for the upload

[CrazyJim1]

50k/sec upload suuuuuuuucks. At least make it 128k/sec like other broadband. The reason I care is that I'm making a P2P video game, and I have to go with the slowest upload for broadband: Comcast's 50k/sec upload. If it went to 128k/sec upload, my game could support 2.5x as many players!

Re:Too bad Comcast doesn't do much for the upload

[j79zlr]

I have the standard package from Comcast here in Chicago. My upload speed is ~512kbps and download bandwith has increased twice in the last year up to ~10Mbps steady and upwards of ~20Mbps burst.

Re:Too bad Comcast doesn't do much for the upload

[Anonymous Coward]

Actually, if it's truly peer to peer, then you run into the problem with the network effect. Basically, as you add clients to the system, the amount of data traffic is going to scale with the square of the number of nodes, not in direct proportion.

So you'd only be able to support 60% more clients (1.6x) with 156% more bandwidth (2.56x).

If you don't use a true peer to peer system, then you can reduce the load by dynamically setting up a hierarchy. Doing so is a non-trivial problem, however; it's basically a multicast problem, which doesn't lend itself well to interactive applications like games. Even if you manage to lick that issue, you still have potential problems with cheating by the privileged nodes in the system (man in the middle attacks).

I doubt that you're doing this as anything more than a hobby, maybe with grandiose plans for the future, so unless you want to buy a third party solution, I think the modest 60% increase is probably more in line with what you'd be able to achieve. A properly-designed game should easily be able to fit its communication into a 50 kbps upstream, though, at least for a reasonable number of players; games don't need a lot of precision (spatial or temporal), which tends to eat up most of those bytes in naive implementations. You may also want to think about doing a server-based system.

Re:Too bad Comcast doesn't do much for the upload

[Anonymous Coward]

Oh, by the way. Since this is an n^2 process, the diminishing returns can get quite extreme. I have a 768 kbps up connection myself (courtesy of the former SBC), which is 15.36x faster than 50 kbps, but would only allow for 3.9x as many clients. Also, this is assuming the packet size remains the same, which probably isn't true as you add on clients.

Mysterious speeds

[suv4x4]

The results surprised me. Here are the reported download speeds in Kbps: New York, 18,946; Washington, 15,821; Atlanta, 11,257; Chicago, 10,042; San Francisco, 4,230. What is going on? I know my father is not paying for a 10+ Mbps connection. Is Comcast giving priority to popular speed-test sites?

Ok... and now try with the cache disabled.

Just kidding...

Seriously though, I think he's probably clicked one of those "Click here to make your internet faster" ads... I knew it they work!

How fast to download the Linux Kernel?

[davidwr]

You probably already downloaded and built today's patch, but just for grins go download the whole file.

What, you didn't download the patch yet? What kind of geek are you? :)

Re:How fast to download the Linux Kernel?

[JimDaGeek]

The 2.6.22 patch came down at around 5.5 Mbps. Which is weird, because the speed test sites have generally been accurate from my TW cable account down south. However, the speed test here still shows 12+ Mbps, though as I pointed out, big downloads are at or near the 6 Mbps mark.

Why is this even being posted?

[Anonymous Coward]

Those are burstable rates, not sustained. The fact that this is even being posted on slashdot based on this ONE guys SINGLE speedtest using powerboost is kinda funny, shame on kdawson. Ask anyone in the DSL Reports forums, this is a joke if anyone is to take this seriously.

http://www.dslreports.com/forum/comcast

This is also an ad for Comcast, "OMG CRAZY FAST SPEEDS FOR SAME PRICE AS SLOWER SPEEDS!!", you just dont know it. Say basically, in one day, we have one post slamming Verizon for taking residents copper away and another praising Comcast for super fast speeds (Take that FIOS!!). Doesn't seem strange to anyone else?

Do It Yourself

[Doc Ruby]

The Speakeasy speed tests are indeed easy, and easy to "speak" about on their site with posted ratings. But there's nothing magic about it, that you couldn't do with simple commands from your PC.

All you've got to do is fire up a shell (whether Windows, Linux, or other client OS), and download a big (>10MB) file while timing it. Find an HTML link to a video or something, then download it from the shell (eg. wget or curl in Linux) to a local directory. Watch the minutes and seconds from when you first connect (right after you give the command, after you get the download feedback), to when the file is complete. Then examine (eg. ls on Linux, or use your GUI file manager) the file for its exact size in bytes, then divide the time by the size.

I know this seems obvious, but distrusting Speakeasy's numbers as cooked by Comcast shouldn't be the last act before punting to Slashdot. Real tests, not just examples like Speakeasy, are trivial to run by yourself.

Re:Do It Yourself

[tomhudson]

Under linux: time wget "url_of_file_you_want_to_time.iso"

Why do the math when there's a program ready to do it for you?

Re:Do It Yourself

[Solra Bizna]

Unless you've found a server whose bandwidth is lower than yours.

Also, 10MB is not big enough for a reliable test. I'll arbitrarily say 200MB minimum for a fair test.

-:sigma.SB

Re:Do It Yourself

[Doc Ruby]

You don't need "time": wget will report the average DL rate when it's done. But since they asked such a basic question, I felt like being pedantic.

Re:Do It Yourself

[Anonymous Coward]

My experience seems to indicate that Time Warner is doing major traffic shaping in the Tampa area. When our service was first connected, my neighborhood saw more or less typical cable modem speeds on about 4 days per week, and 1/4 or less typical cable speed about 3 days a week. The cycle was irregular but the slow days always came in clusters. I assumed that this was due to user traffic variations and system maintenance issues.

Then, Verizon ran a fiber optic cable down our street. Road Runner cable speeds immediately rose to previously unseen heights, and the slow days are a fading memory. I consider this evidence that we had been experiencing a "rolling brownout" cycle all along, and that we were moved from this rotation to a higher tier, in response to the appearance of a competing service option in our neighborhood.

Small wonder the ISPs adamantly oppose the collection and publication of detailed statistics on locations and speeds of "broadband internet service".

Re:Do It Yourself

[Doc Ruby]

On TWC RoadRunner here in Brooklyn, I get 10Mbps downloads, and something like 0.5Mbps uploads for my $50:mo. They keep increasing the speeds without any announcements. It's kinda weird. But I hear that OptimiumOnline in the Bronx gets 20Mbps/2-3Mbps, paying something like $30:mo.

Re:Do It Yourself

[Anonymous Coward]

opti is also 50/month .. 45 if you have video .. that 29.99 a month deal is only for new customers .. and even then only for the 1st year .. Howevere I did see a comericial saying that 'now existing customers can get' that 29.99 deal ..

Oh .. also the stated speed is 15 down .. with 30 is you pay extra for the "boost" service ..

test with akamai network

[Anonymous Coward]

use: curl -o /dev/null http://images.apple.com/movies/wb/300/300-tlr2_h10 80p.mov
or any other trailer there
flash based speed testing sucks.

Initial speed bursts

[DreadfulGrape]

I have Comcast 8 MB/sec. service. When I begin a long, bandwidth-intensive operation my throughput will start out in the low teens, like 13 or 14 MB/sec. Then in about five seconds it will throttle back to just under 8 MB.

Perhaps something like this accounts for the Speakeasy.net results.

Caching?

[beaverbrother]

Perhaps the speed tests are used frequently enough for their system to cache it automatically?
As long as it isn't special treatment for the test sites, it is hard to blame Comcast for the speed boost.

Comcast speed

[Anonymous Coward]

I did have Comcast up until about 6mo ago when I moved to a Charter area. I paid for an 8mbit connection, and would get bursts much higher than this, it is due to their power boost, or whatever it is called. Not just on speed tests, I would get sometimes as high as 16-20mbit downloading from bit torrent, for periods of time. There is some restriction like 'first 100megabytes of a download' or something similar, not exactly sure what the restriction is, but the connection does not continually run that high all the time. It is not any kind of web caching like is used in some dial up speed boost technologies, they actually do raise the bandwidth cap. Currently I have Charter cable internet with a 10mbit connection, and since replacing all my inside wiring with rg6 quad shield, I get 10mbit down/1mbit up consistently. Charter is also starting to launch 16mbit/2mbit in parts of texas, which I think is very exciting. To the poster stating that he gets slower speeds in LA during certain times, there is most likely some other problem causing this, I used to work High Speed Data support for Charter, and I have seen the bandwidth usage for the uBRs in the area, and they do not come close to meeting the available bandwidth. Charter is very good about balancing their customers and they do not oversell areas, if customer base increases to stress the bandwidth for the area, the area gets upgraded to meet those needs. Currently the cable modem standard is Docsis 2.0, I was reading about the development of Docsis 3.0 which uses channel bonding to provide much greater speeds, recently Arris demonstrated a channel bonded Docsis 3.0 modem that ran at 150mbit/sec, and Cisco one that ran as high as 298mbit/sec. I feel this will be an awesome technology to help drive the internet bandwidth market and provide the competition that FiOS needs to increase bandwidth and lower costs, hopefully America can move ahead in the broadband market to bring us back up with the rest of the world.

Lots of factors...

[SuperBanana]

To name a few:

• Router/signal compatibility. Comcast has switched signals several times and in two cases, it caused horrendous performance. There's also a particular cablemodem (has a big blue "sleep" button on top) that has serious problems with high packet rates and connections (ie, BitTorrent crashes it.)
• Interference.
• Local loop, backbone, and uplink utilization. Guess what, guys? Sometimes traffic peaks for strange reasons. Sometimes it's a virus outbreak, a new movie trailer, or a big news story.
• THE INTERNET . It's unreliable, not guaranteed, never has been, and YOU ALL KNOW THIS AND HAVE BEEN TOLD IT OVER AND OVER AND OVER. I work for a university. We have seriously fat pipes. We have a 10Gbit backbone. And some sites I can FTP from at 2MB/sec. Others, I get 40KB/sec. "Speed tester" services compete for bandwidth just like everyone else. Stop holding them up as some pure, holy source of internet traffic that magically flows through every router at maximum speed.
• Some content is akamai-zed. I get Apple's software updates at maximum line speed, for example.

If you're not happy with your service, CALL THEM. My parents were some of the first people to get MediaOne service back around '98-'99, and every time they had problems, we picked up the phone, and it was taken care of.

I've had the same experience elsewhere. Any time I have problem with the service, be it regular disconnects or lousy performance- I pick up the phone, and a few minutes later someone is checking into signal to noise ratios and such. If you lease the modem, they're usually happy to try sending out a tech and swapping out a modem if you're polite but clear there's a problem. They're usually even more amenable if you pick up the modem yourself at a "service center."

In my years as a customer and having friends who were customers, I've seen a)flooded junction boxes b)in-house distribution amps turned up too high c)1 failed modem d)one buggy model e)several incompatible modems after "upgrades" to the area network (usually to support faster speeds.)

In short: call comcast, ask them to look into it. They've almost always been helpful, through all the various company changes: MediaOne, RoadRunner, etc.

Re:Lots of factors...

[pokerdad]

In short: call comcast, ask them to look into it. They've almost always been helpful, through all the various company changes: MediaOne, RoadRunner, etc.

As someone who used to work for Comcast allow me to say rotflmao. Either you are one lucky sob or you are lying.

Just as a matter of example (one among many) during the entire nine months I worked for Comcast the entire state of Illinois never left the outage board. That isn't to say that no one in Illinois ever had a connection, but many people had little or no connection and we were under instruction to do absolutely nothing for anyone from Illinois - just keep BSing them till they gave up.

Re:Lots of factors...

[Anonymous Coward]

in short, your a moron, what you are discribing, line noise, and signal, are, unless the cable system is 40 years old, which most arnt, irrelevent. DSL relys on these things. you dont work for a university, because 10gbit and your getting 2MB/s is retarded. Stop ranting and read before you post nonsense

Re:Lots of factors...

[entrigant]

You should check your eyeglass prescription. The article is inquiring about how he got HIGHER than advertised speeds. Your reply seems to be answering a different question...

Re:Lots of factors...

[Urza9814]

"There's also a particular cablemodem (has a big blue "sleep" button on top) that has serious problems with high packet rates and connections (ie, BitTorrent crashes it.)"

There's a lot more than one. I've gone through two with that exact problem myself. Neither had the big blue sleep button. And it was pretty much hell to get them to replace it...called, they sent a guy to check it, he said nothing was wrong. I had to dig around online and pull up a few references showing that it was a known problem with that model of modem before they'd replace it. I have a feeling they just took the old one and gave it to someone else.

Re:Lots of factors...

[Danga]

I am curious if the time you were at Comcast was last year at all because I used to live in the Chicago area the last couple of years and I had Comcast cable TV and internet?

Well last year was HORRIBLE as far as my internet connection provided by Comcast was concerned. It got even worse around May (it always seemed the hotter it was outside the worse my connection became) and never got any better. Basically my cable modem would keep randomly disconnecting for 5-15+ minutes at a time (it was replaced more than once) and if I remember correctly the logs on it seemed to say it had a problem syncing and the SNR was bad too. This was all after 2 years of everything working without a hitch. I gave up on WoW and any other online game that needed a low latency connection and the most annoying thing was getting disconnected in the middle of the game due to the connection dropping completely. The only thing I could still somewhat play was online poker but near the end I gave up on that too due to getting disconnected at key points in the game and not being able to get back in in time since it took so long for the connection to go back up. I luckily was able to get my money refunded for a few months but it really was a pain in the ass.

Now I am in Phoenix, Arizona and have Cox cable and while the internet has been MUCH more reliable (and actually usable!) I must say the cable TV sucks ass. The guide is a piece of shit and I don't use it at all since it is so worthless, has horrible search capabilities (you can't even search by genre), the HD channels will randomly go out of service for minutes at a time, and there is no On Demand. The On Demand Comcast had was one thing I loved, especially the amount of free content. Anyway, before I moved I thought I would never miss Comcast at all but after experiencing the Cox Cable TV service I actually do miss Comcasts cable TV!

Some more info

[JimDaGeek]

Testing locations closer to Philadelphia gives the highest speeds. However, other locations seem to vary greatly. For the curious, my father pays $45 a month for his cable modem broadband.

Using the speedtest.net site that was posted above, I am getting all kinds of different results.

Toronto: 2,790 kbps down, 1,216 kbps up<br>
San Jose, CA: 2,101 kbps down, 470 kbps up, 92 ms ping<br>
New York: 17,410 kbps down, 1,357 kbps up, 12 ms ping<br>
Tucson, AZ: 4,258 kbps down, 1,461 kbps up, 190 ms ping<br>
Mexico City: 2,811 kbps down, 992 kbps up, 81 ms ping

Down in Columbia, SC I generally get consistant speeds of 8 Mbps. Comcast's speed boost thing does seem like a nice bonus, though would this cause any other locations to slow down? Less than 4 Mbps to CA is a little slow IMO.

Re:Some more info

[Mezoth]

The higher your latency to a site (aka, how long does it actually take a packet to get there) the slower your bandwidth will be without some extensive TCP window tuning. Reference:

http://en.wikipedia.org/wiki/TCP_Tuning

That is why you get slower speeds the farther out you go, even assuming every link in between is not congested.

  Also, most of what you post there is explained by powerboost - very fast downloads for the first few seconds, then they throttle it back. If you ran a sustained download (go get a linux distro!) you could see what the actual sustained download speeds are after about 5 minutes.

Re:Some more info

[JimDaGeek]

Yeah, I know about latency and all that. I just never noticed it as erratic as I have here on Comcast's network. Before today I have never heard of the powerboost thingy. Seems a little gimicky to me.

Re:Some more info

[Mezoth]

Yeah, Powerboost is a gimmick as far as I can tell - it gives you a few seconds of very fast downloads. But for the majority of users, this is perfect - those pictures come down really fast from your email, or that youtube video you were linked loads up in moments, etc.

  As for erratic, you are going cross country with most of those results. I would expect a good amount of erratic behavior, especially since you are also dealing with whatever load the speakeasy server has at that moment. The general trend of the results you posted was : the higher the ping time, the slower your maximum speed (Upload was affected far less, which really tells the tale). It is what I would expect, and the noticeable outlier of Tuscon is probably more of a site load issue when the test did the ping. The speed of light is fairly immutable in these matters!

20 Meg Cable Modems Practical Since 1999.

[Dave Burstein]

19 megabits on an unloaded cable modem node is quite practical, especially for brief speed tests.The U.S. cable modem architecture is based around nodes that share 35 megabits or so downstream. RCN even sells a 20 megabit service, which David Reed buys and reports gives him the promised 20 megabits most of the time. In practice, most nodes run with 5-10 megabits typical load, as only a fraction of the time multiple users on one node simultaneously download megabits and fill the pipe. Statistical multiplexing (sharing) works much better than people expect on cable modem networks, especially on well managed nets that split nodes and otherwise expand capacity when needed because a node gets frequently congested.

Comcast and cable suppliers are working on several techniques to allow customers to get more than the 6 or 8 meg typically allocated, while not causing undue congestion. "Speed burst" technology tests the network load, and if uncongested temporarily raises the speed of an individual modem making a fast download for a brief period. While that's marketed as "doubling" speeds to 12 and 16 megabits, bursts to 20 and 25 megabits are also practical.

The new technologies require upgraded equipment and are typically being tested first and then rolled out market by market. So it would be no surprise if a subscriber in Philadelphia (Comcast's home town) is benefiting from a test or early deployment of faster speeds than Comcast customers elsewhere.

100 megabit+ (shared) cable modems are being deployed in Japan, Quebec, and France, bonding 3 or 4 35 megabit channels for higher speeds. These are early "DOCSIS 3.0" products, unlikely to be widely deployed in the U.S. until 2009. Comcast's CEO, Brian Roberts, demonstrated 100 megabits at the cable show in Las Vegas this spring, and will probably test widely in 2008 and go into deployment (especially where Verizon is building FIOS) the following year. DOCSIS 3.0 requires a new cable modem unit, however, so this customer is unlikely to be an early tester.

That doesn't explain why the test to San Francisco only ran at 4 megabits, which could be explained by node congestion a few minutes later, inferior Comcast backbone connections to Speakeasy's host in San Francisco, or other circumstances. For more details on coming faster cable modems, google DOCSIS 3.0.

Dave Burstein

Editor, DSL Prime

Re:20 Meg Cable Modems Practical Since 1999.

[Mezoth]

The slower speeds to the west coast sites (Phoenix, San Fran, Mexico City) are all explained by TCP window issues.

  Remember, the higher your round-trip time the slower TCP will let you send bandwidth unless you tune the TCP window upwards. For reference: http://en.wikipedia.org/wiki/TCP_Tuning

  This happens with every ISP, and quite frankly when you start talking cross country distances the speed of light becomes a factor in how fast you can actually get round trip time down to.

Re:20 Meg Cable Modems Practical Since 1999.

[DNite]

Amen. I was surprised that a mention of Comcast's advertised SpeedBooster technology fell this far down the thread. I can verify that Comcast does boost the speed of downloads (and sometimes it seems even uploads at times) for a brief period of time. I think it's somewhere around the first 20MB or so get a nice boost and then your download tapers down to the speed you should be getting. So technically Comcast is going to get good marks in the speed test because of the brief boost you get, but this boost happens on nearly all downloads where the sending server can manage that speed...

Upload blows

[DTemp]

I don't care if my download speed is 50mbps as long as my upload speed is still below 3mbps (currently its 768kbps).

I log into a VPN every day and upload a few dozen 10-megapixel 16-bit CMYK TIFFs, averaging around 80MB each. It takes about 15-20 minutes each. Thats close to 8 hours a day saturating my upstream connection.

If they would just give me 5mbps upload, it would take only 2-3 minutes each, or a little over an hour.

I'm on Cox...

[networkzombie]

where I get my as-advertised 6Mbit for $38 (not including fees). The 12Mbit is available, but it's $54 (not including fees). I figure I'll offer to pay for the 12Mbit plan when they threaten to cut me off for exceeding my bandwidth. One HD movie at 720p can be 4 or 8 gigs. The 1080i stuff is large.

Re:I'm on Cox...

[Anonymous Coward]

It's funny, because I downloaded like half a terabyte last month, and got no notice from comcast, though I expected one. I'm also getting ~500KBps downloads in emule at times, with no blocked ports. Awesome service imo. Although I really had it made before I moved from across town, there I was getting 24Mbps out of their [well, technically, RR] 5 Mbps service.

So encrypt your P2P traffic...

[Anonymous Coward]

... and refuse non-encrypted connections. If they don't know what it is, they can't shape it. :)

Re:So encrypt your P2P traffic...

[Fatal67]

Encrypted P2P still has a fingerprint. While they don't know what is in the file, they do know it is P2P and it will be rate-shaped along with the rest of the p2p.

Encryption would work on foiling AT&T's filtering technology. They can't decrypt a flow on the fly without the NSA's secret pass key. The NSA wouldn't give that to AT&T as AT&T has not been very good at keeping secrets lately.

They are up to it.

[WindBourne]

I tested my comcast 6 MB with and got 1.2 MB., but With speakeasy, it ranges from 8-12 download. It is obvious that they are selectively caching.

Re:They are up to it.

[Technician]

Maybe not.

I tested my comcast 6 MB with and got 1.2 MB., but With speakeasy, it ranges from 8-12 download. It is obvious that they are selectively caching. [bandwidthplace.com]

Who are you connecting to? For a while, I had a 60 Meg connection at work. DSL speed tests witht the graph bar were fun to print out an post. With a speedy connection at work, I could easly find ISP's and webhosts with serious connection problems. Many things pop up like they are on my local hard drive. Other things load like dial-up. During the Napster heyday, most anything .edu was guranteed slow. (not Napster transfers, but school webpages. I never did Napster) Since those days my connection has slowed as there is more filtering at the proxy to slow things down so it's much harder to identify ISP's with slow connections.

When I first got the job, I surfed ISP websites to find an ISP with a good net connection to use from home. What good is an ISP with a speedy connection to the ISP and it's small garden when it was severely bottlenecked to the internet? I eventualy found a good local ISP with a great connection. When testing internet connections, be sure if possible to check your connection tot the ISP as well as to the Internet at large.

In short, If you are with Speakeasy, you might want to also do speed checks to the DSL Reports tests.

Comcast detects speed tests

[Anonymous Coward]

This is more visible with upload. Using the speed test I get 1.2 mbps up, but in actual usage (p2p) it caps out at 40 kb/s = 320 kbps, which is only 25% of the speed test result.

The real problem, however, is not speed but reliability. Comcast service has unscheduled sporadic downtimes, and when they find a problem on their end they pretend that it wasn't really there. For example, the internet service was out for 3 days -- when I finally called them to ask what was going on, the technician said "I don't see anything wrong from here" and then after I hung up, the connection was magically back on.

Also, did I mention that Comcast uses traffic shaping (against p2p)? You have to use obfuscated protocols in order to get around it. By contrast, the SBC DSL I had a year ago was far more reliable, even if peak download speed wasn't as high (not that my downloads on cable go much past 200 kb/s on average anyway due to throttling on the other end), and they didn't try to limit bandwidth usage.

The results are erratic

[Dracos]

I did several consecutive speed tests last year (I'm on Comcast in Sacramento), and the results were erratic at best.

Over the course of about an hour, 20 tests said my bandwidth ranged between 3Mbps and 24Mbps (all to the same server, average was about 9Mbps), and I pay for 8Mbps.

In a word, it's Comcastic!

Chicago Speeds

[dextration]

I live in Evanston IL, and I was pleasantly surprised to see the number '25,619' come out of my download rate from Speakeasy. I have a menubar (macintosh) addon that displays my ethernet throughput, and it topped out at about 3.0MB/Sec. I am quite happy with with Comcast, for once.

I have Comcast...

[LeRaldo]

I live in Washington, and I pay for the 8Mbit package Comcast offers. I get up to 24Mbit download (constant) and bursts of 1.5Mbit upload (when it's not bursting it sits at 768Kbit). I've always gotten more speed than I've payed for since Comcast took over ATTBI in my area. It was only very recently that I started getting such good upload speed though; the package I pay for comes with 768Kbit upload, but they must have changed something to allow for fast bursts.

Re:I have Comcast...

[LeRaldo]

And I forgot to mention that these speeds I'm getting are on torrents, not these speed test sites.

It's probably your tcp window size

[Anonymous Coward]

Notice that the throughput gradually decreases as the physical distance increases. Most likely, your TCP window size isn't large enough to fill the pipe. Do some research on this and try implementing window scaling (RFC 1323).

Not really news

[Anonymous Coward]

Comcast frequently tests out high speeds without telling customers. Esp. in the Philadelphia area (corp. HQ) and Denver (former AT&T Broadband corp. HQ). It'll be gone again in a few days.

-former Philly corp. HQ disgruntled employee posting as anon coward so the big C doesn't sue him/her into oblivion.

Full speed to common speed test sites?

[AHuxley]

From "Analysis: The White Lies ISPs Tell About Broadband Speeds" at
http://www.pcmag.com/article2/0,1895,2155140,00.as p

Page 2 has a small part on:
"3. Does your provider give preferential treatment to speed test sites?"

Mb a few slashdot ppl could map this out and post some results?
Be fun to see if the 'test' car/cab/bus/track gets a free lane.

REAL BANDWITH TEST

[logik3x]

http://miranda.ctd.anl.gov:7123/ Can't stand seeing people using speedtest.net, MOD THIS UP!

Re:REAL BANDWITH TEST

[Revotron]

From the anl.gov test page:

Another client is currently being served, your test will begin within 3915 seconds The government just got /.ed...

Re:REAL BANDWITH TEST

[X0563511]

And further down are links to other systems... most are working

Re:REAL BANDWITH TEST

[Ingenium13]

No, mod this down. At least speedtest.net doesn't make me wait 12195 seconds to do the test and doesn't crash Firefox.

Re:REAL BANDWITH TEST

[s0l0m0n]

15930 seconds, and crashed my browser!

Great app.

Re:REAL BANDWITH TEST

[eat here_get gas]

I tested with FF. from CT to LA I got 6000/169 with 114 latency, yet CT to NY was -3000/149 with 129 latency. I tested to LA as all my bootlegs I DL don't come from next door....

I also had NO wait, and FF never flinched.

Re:REAL BANDWITH TEST

[psydeshow]

That's not Firefox you hear crashing. It's Sun Java.

Took about 30 seconds to fire up the VM there. FF 2.0.0.4 on OSX 10.4.

Re:REAL BANDWITH TEST

[Kamokazi]

I think THEY just got a bandwidth test...

Who cares:

[information_storage]

The answer is: Who cares!? That's 12mbps burst by the way, and the bursts don't last for nearly as long as they say they do. In addition, who needs over even 5mbps (which is higher than what I get with comcast) if you still have near-56k upload speads. And yes, "386k/s uprate" is rediculously low in comparison to what we should be getting for over 40/month as I pay in the monopoly ridden area of Chicago. In short, 5GB/s downrate is useless without any uprate, and that is how they choke us (that and service that goes out all of the time). In the end, 99 percent of the user base is too nieve to notice; as long as the people stay as tamed and fickle as they are in the US, my connection is going to suck.

Adelphia to Comtrash

[BanjoBob]

When Comtrash acquired Adelphia, my net speed dropped from 6 Mbps to 1 Mbps. I called them repeatedly and they came to the house repeatedly to test. Yep, 1 Mbps was all I could get. But, with Adelphia, it was a reliable 6 Mbps. Finally a service rep said that most people don't know how to test their connections and that Comtrash was reserving the bandwidth for their new VoIP and telephony products. So, I dropped my $60 per month 1 Mbps and with with a $30 per month 1.5 Mbps. I got 50% more for half the price!

Re:Adelphia to Comtrash

[u-235-sentinel]

When Comtrash acquired Adelphia, my net speed dropped from 6 Mbps to 1 Mbps. I called them repeatedly and they came to the house repeatedly to test. Yep, 1 Mbps was all I could get. But, with Adelphia, it was a reliable 6 Mbps. Finally a service rep said that most people don't know how to test their connections and that Comtrash was reserving the bandwidth for their new VoIP and telephony products. So, I dropped my $60 per month 1 Mbps and with with a $30 per month 1.5 Mbps. I got 50% more for half the price!

I feel for you. I was a Comcast customer for nearly 4 years and couldn't believe the games they pulled. We were even looking at replacing our TV because the signal kept getting worse and worse. They said we needed to replace our TV so we had nearly $2000 budgeted. What's interesting is after they terminated our Internet service and started playing word games with AUP garbage we switched to Dish and discovered our TV was just fine. It was them all along!

Oh and yes, they disconnected us for using the Internet too much. But of course they won't say what is acceptable. Only what isn't. Very strange company.

rather nice if you really have that

[r00t]

Lots of places get many MB up, but only 128 kB down! This is no good for videoconferencing. The 128 kB is bad enough to affect downloads even, because TCP ACK packets can't get back fast enough.

374 KB, especially if "rock solid" (no VoIP dropouts unless you go above, low latency, reliable 24x7 operation) is great.

Re:rather nice if you really have that

[gelfling]

I suppose you're right & I shouldn't complain. The problem(s) is:

I'm a permanent work from home'r so working on a network that assumes 10x higher bandwidth TO me instead of from me is problematic. Sending business sized presentations objects around is pretty damn slow. The asymmetric nature of it, is killing me.

I have VoIP - from AT&T Callvantage. I'm not sure if the problems are bandwidth related but the service is ludicrously unreliable. My employer is spending $25/month to get me a service I leave on the side unused 90% of the time in lieu of my own cell phone.

But let's not ignore that the US is a decade behind the rest of the world in terms of broadband. And there's no end in sight.

Re:rather nice if you really have that

[clemclone]

Let's see here: Raleigh, NC, works from home, uses AT&T Callvantage -- must work for IBM. If so, don't worry, they'll probably be laying you off shortly.

Re:rather nice if you really have that

[gelfling]

You're almost right but you left the 6-10 split.

not compression

[r00t]

Sending the bits is easier than compressing them. Compression is not as simple; it takes compute power and time.

Probably they just allocate an extra (second or third) channel to you on demand, maybe with something to keep you from having it continuously.

Re:not compression

[fm6]

True, but it isn't that hard to add compression hardware. Every modem (back when "modem" still meant the modulator-demodulator you used to send digital data over an analog circuit, instead of the digital network bridges that replaced them) used to come with a data compression circuit. It's the only way you can get 53kbs throughput on a voice line.

I don't actually know why internet data packets aren't routinely compressed. But I suspect that it has more to do with making the protocols dangerously complex than with the (minimal) overhead of data compression.

Speed tests, DSL and Cable

[mrslacker]

I'm in SoCal, and there's really only two choices - AT&T DSL (1.5, 3, 6Mb) or TWC (~5Mb). I had 3Mb at home, and it was almost always possible to max that out. I recommitted for 12 months to 6Mb (worked out being about the same price). That takes a while to ramp up, and doesn't always max out - of course, AT&T's claim is "at least 3Mb".

Having said that, at work we have 3 x 6Mb business DSL lines, and it's easy to quickly max those out, although it can depend on where you're getting stuff from. mirrors.kernel.org for example tends to be busy at times, and the Debian ISO mirror is in Sweden (I think). From what I hear about cable, it tends to be over subscribed around here.

So why do we have 3 you might ask? Well, we don't - we actually have 5 separate lines. I won't say too much about the details to avoid this being an advertisement and because it's not quite ready yet, but we have a product that aggregates broadband connections using a variety of tricks and techniques, giving you essentially linear increases in speed for every connection added (for HTTP anyway, other protocols are load balanced). The SpeakEasy test reports north of 16Mb/sec. For large file downloads, (HTTP and bittorrent, depending) you can get over 1.5MB/sec. Other broadband speed tests may report bogus values because they try various tricks and our setup confuses them, YMMV. (And yes, before anyone starts, we understand all about multiple IPs and authentication, etc, etc).

Of course, the available speeds in mainland Europe (and some parts of the UK, although there's a huge range of variation there) tend to be much higher. I'm stuck in backwards California.

sounds like channel allocation

[r00t]

Cable modems are strongly based on TV channels. There are separate channels for upload and download. Normally, you share a channel with other people. An upload channel is time sliced; everybody gets a turn to transmit. (like GSM phones) A download channel shoves packets for everybody, a bit like switched ethernet.

There are extra channels. When you have a really big download, the cable system will give you exclusive use of one of these extra channels if there is one available. You thus get really high speed. At the very beginning of your big download, or when there are none of these extra channels available, you have to share.

Some places have lots of these extra channels. Some places might have none. Channels for this purpose take away from other things, like pay-per-view and home shopping.

LanSpeed

[BanjoBob]

There is a nice little freeware program called LanSpeed by OrcaWare. I don't know how reliable the numbers are but its fun to watch your uploads and downloads.

I remember that when I had Comcast, my network traffic was pretty constant -- even if I wasn't doing anything! When I switched to a DSL line, that all stopped. I never did figure out what all the Comcast traffic was about.

Anyway, you can find LanSpeed at
http://members.chello.nl/~m.vanosta/orcasoft/index .htm It is a fun little utility but, again, I don't know how accurate it is. The relative speeds are close to what I'm seeing.

Enjoy!

I just want lower cost

[DaSH Alpha]

I could care less about Comcast's Uber Super Dooper Faster Than Your Turtle DSL Power Boost Speed. Most usage of the internet can't even take advantage of the speeds they supposedly offer. All I know is that if they offered a slower speed for less $$$ I'd switch in a heartbeat. Although for the moment I'm just thanking God that they've never increased their internet prices (bastards just raised cable by $3 again).

Re:I just want lower cost

[seahawk2199]

There is a slower speed tier. At least in the Minnesota region. It's only used for retention purposes and not advertised. They call it the Economy Tier, I think it's around 768 kbps down and runs $30.

Re:I just want lower cost

[Greg01851]

in MA there is an economy plan. $25/mo for 384kbps upload/download. Not sure why they skimp so much on the download and keep the upload speed the same as regular plans, but there it is. Ok if you just do some email and surf, but not so good if you have VOIP or do a lot of downloading. It's an unadvertised plan, you have to call them to get it. I saw it when they did a mailing about some rate changes, and listed all their plans.

peak vs. sustained throughput

[davidwr]

Getting high peak bandwidth is a nice bonus, but what counts when you are watching live video or what not is sustained throughput.

That's what you are paying for.

If you are paying for 5 and getting 5.5, great.

If you are paying for 6 and getting only 5.5, time to call the board of advertising regulation.

Re:peak vs. sustained throughput

[JimDaGeek]

If you are paying for 6 and getting only 5.5, time to call the board of advertising regulation.

I think the ISP's have all handled this in their EULA. I remember something in mine about no guarantees. Must be nice to be a company that sells a service with no guarantees of performance, reliability. If I get 1 Mbps instead of the 8 Mbps that I pay for, well they just blame it on the "Tubes". ;-)

18000kbps? try 25223

[clamothe]

http://image.bayimg.com/ca/dh/da/ab/i.jpg Location: Eugene, Oregon. Internet: Comcast.. just the regular residential plan.

Re:18000kbps? try 25223

[armanox]

Damn. That beats my average on the school's T3.....

Comcast is fast...

[Newer Guy]

My brother (who is in the Boston area) has their standard service. He's been getting 15,000/3000 Internet for quite some time now. It even bursts to almost 30,000. he's getting amazingly fast speeds. I saw download speeds on his computer well in excess of 2 mB/sec. I wish I had those speeds here in L.A. with Road Runner. On a good day I get 9600/960 and I have their extreme service (which costs an extra 10 dollars a month).

Who cares about peak rates?

[ivan256]

My problem with Comcast was never how fast it could be. It was the frequency at which it wasn't working at all.

When they work on reliability, up-stream, and price, then maybe they'll deserve some "props". For now, Fios is kicking their ass up and down the block. Even these theoretical burst rates can't compete, plus Fios is cheaper and more reliable. Sorry Comcast.

amazing

[Dahamma]

The post is basically about how Comcast, against all previous expectations based on their crappy service, has actually provided MORE than promised? And that makes it time to bash them again?

Hey, I can't stand Comcast as much as the next poor customer (probably more, since I used to work at @Home, which was killed by these damn cable companies) but give be a break, they have actually done something that benefits the average customer (give them huge initial download bandwidth to make web browsing FAST, but excessive P2P, etc, traffic only "as advertised"). And that is interpreted by many here as "manipulating the bandwidth tests"?? Yeah, I'm sure Comcast also helped fake the Moon landings.

Re:amazing

[1310nm]

In my area, @Home was offering Internet service over my local cable company (then TCI), and when AT&T bought TCI, they called it AT&T@Home (assuming @Home was also bought by AT&T?). Once AT&T sold the franchise to Cox, it was simply Cox HSI. Curious how that worked out, did @Home actually compete with newly built cable networking infrastructure shortly, or was it acquired in pieces by the franchises?

Re:amazing

[Dahamma]

Actually, @Home *built* all of the cable companies' network infrastructures and provided the broadband service for pretty much every major cable co except Time Warner. They had about 4.5M customers paying $40+ a month (which was about half of the total US broadband market at the time!) Can't lose, you'd think? Well, the catch is said cable co's also owned a significant stake in @Home, and once they decided they could make more money doing it themselves, they let it die...

Here's a decent post mortem (sounds surprisingly like "Who Killed the Electric Car?...):

http://sfgate.com/cgi-bin/article.cgi?file=/chroni cle/archive/2001/12/17/BU23049.DTL

And I think agreeing to pay $340M to the @Home bondholders is as good as AT&T admitting blame...

http://www.comcast.com/About/PressRelease/PressRel easeDetail.ashx?PRID=152

I was so excited to see AT&T all but disappear as a public company... until SBC goes and builds a new Deathstar three times the size of the original...

hmm.....

[Sillygates]

Maybe they are just getting priority because the rest of their time on the net is browsing, Comcast could have detected your large downloads and enabled their special boost mode service:
http://www.dslreports.com/shownews/75298

In Bay Area try Apple.com

[toonerh]

They have lots of 10's to 100's of MB downloads that are compressed already, so provider compression won't help. Of course if the file is in a squid cache, nothing upstream matters. (download Mac OS X 10.4.3, not iTunes 7.3).

Speakeasy vs Stanford's speed tool

[mishikal]

I have Comcast for my internet provider, so I did a test using Stanford's speed testing system as well. Speakeasy reported: 10,759kbps down 1,766kbps up Stanford reports: running 10s outbound test (client-to-server [C2S]) . . . . . 1.78Mb/s running 10s inbound test (server-to-client [S2C]) . . . . . . 6.55Mb/s So the up speed is virtually identical, but the downspeed is a bit slower. Although the routes between my house and Stanford may be a bit different between it and Speakeasy, I'm surprised at the rather large difference. These two tests were done within a minute of one another.

On Cable depends how many people use it...

[Taelron]

I live in an appartment complex that refused to install a $1 million dollar repeater to support cable broadband a few years back. So most of the people living here have DSL only because of a deal between the phone company and the appartment complex that locks you into one provider. Ironically, my building is the point the cable system enters the entire complex (10 buildings). So only people in my building can get cable modems according to the cable companies installers. So even though I am only paying for the basic 384kbs up/ 6mbs down, my actually speeds are typically 30mb down, 2mb up tested to several different test sites. So I'm happy with it. The downside will be when I finally move and suddenly I am down to the same speeds as everyone else... :(

When it was @Home running it speeds were locked at 128k up, 3mb down. When AT&T took over they raised it to 192k up but left it at 3mb down. Comcast took over from AT&T and raised down to 5 mb, and last year is when it appears they removed any artifical locks and speeds increased. On the flipside, Comcast throttles P2P traffic and slows down any P2P connections to the point that programs like Limewire are just painful to use.

I'm just trying to get service at all.

[Anonymous Coward]

I'm negotiating a lease on a house overlooking the Pacific Ocean, at one of the famous surfing beaches. Floor to ceiling glass windows on the ocean side. But there's a problem. The house was purchased in a foreclosure, and the utilities are a mess.

• For phone, the drop to the house has been cut and there's no telco connection from the pole. But it's within range of the CO (well, a substation) for AT&T DSL.
• The house has never had cable. Comcast service is available, but I'd only want data. Does Comcast give you a hard time if you want raw Internet without television or VoIP?
• Sprint PCS - 0 to 1 bars.
• The electric service is rated 15A @ 240V, which is dinky by modern standards.

Incidentally, can anyone recommend a corrosion-resistant PC?

power boost

[DragonTHC]

it's a burstable connection.

Comcast figured out that it's a much better thing to finish large downloads faster to alleviate net congestion.

Their network isn't bad, I have comcast and it's been decent in the past year.

SuddenLink

[MrCopilot]

Rural KY, I have only two options. SuddenLink and DSL. 3mb/s or 1.5. I chose 3. Only I recieved 1Mbs. I use linux and knemo gives me a steady stream of data as to throughput and I have NEVER got even close to 75% of my advertised 3mb, Complaints to Suddenlink and Hours on the telephone has led me to the conclusion that I should have gone DSL. They have since "upped my speed to 3mbs but I havent noticed it. Average DL speed is a pretty constant 31kb/s Less than Dial up for $DIETY sake. This is why I pay for High Speed Broadband?

Comming from Florida and TimeWarner/Brighthouse where I had a constant 10mb/s, Dissatisfied is an understatement.

apt-get update is the most painful part of my existence now. Second only to uploading work back to FL.

Value Propositions

[IBitOBear]

Comcast has no real interest in preventing faster _download_ speeds than what is advertised if there is no line contention. It would cost more than it was worth to truely throttle data going from them to you. They _do_ have a real interest in throttling their upload speed. Consequently on a really idle link there is no real reason _not_ keep the customer-directed data bandwidth full.

In quality of service terms, you put allocations in the QOS table but you let lower tiers "borrow" from higher tiers when the higher tiers are idle.

Now I have tested and _tuned_ my Comcast service and I _do_ get 8Mbps downstream during peak hours using raw linux measurement (so I know there is no compression tricks etc involved), but I had to QOS tune my outbound path to make sure I didn't drop my ACKs. Relaxing (increasing allow data rates) my outbound throttle by 1% will crash the link from 8Mbps to a bouncy range of 3 to 5 Mbps. Really, the outbound throttling to keep the modem from trashing outbound ACKs (and thereby preventing unnecessary fallbacks and retransmissions) is the key to a good throughput.

Someone above was talking about having bunches of separate channels to do a true test. That's pretty clueless. Since TCP scales up linearly and scales back exponentially, spamming a link with lots of diverse TCP data is not usually a stable measure of that link. One stream with a known-stable path, and a known non-saturated return path, is a more correct measure. Of course you have to know how to read the results to understand when you are measuring the link and when you are measuring "the system" (including both server and client and the other links in the transit etc).

So anyway, I have no idea how Comcast implements the "speed boost" thing. They don't even give hints. I can imagine several different ways such a thing could be managed, but they don't provide any technical information to help me make an educated guess.

I have found Comcast's raw service to be quite stable and reliable, more-so now that I have taken responsibility for the upside link. I can imagine that if you hooked a Windows box directly to the cable modem you would get crappy unstable throughput. With no outbound throttle you have the Windows TCP stack slamming that link and hemmoraging ACK packets.

Re:Value Propositions

[JustNiz]

>> Comcast has no real interest in preventing faster _download_ speeds than what is advertised if there is no line contention

Sure they do, they want you to feel the pain so you pay extra for their higher bandwidth service.

Raising speeds due to FIOS competition

[Isaac-Lew]

Personally, I think they're raising speeds in certain areas due to competition from FIOS. When I lived in Loudoun County VA last year, I got 16/2 from Adelphia/Comcast at the same time Verizon was rolling out fiber cables. Now that I'm in PG County MD & I'm seeing orange-colored fiber cables around my neighborhood, I'm hoping for a speed increase here.

I did well w/Comcast

[Down8]

http://farm2.static.flickr.com/1112/759201977_cec6 fe341d_o.jpg

I managed to get a 20M/2M connex, and I believe the last time we received a note about our "upgraded speed" we were rated around 6M/768k. I'd trade it for my 6M/768k and Comcast not capping my damn connection when I'm downloading a lot of torrents. Bastards.

-bZj

A few datapoints

[Francis]

I'm on Comcast, in Portland OR. I pay for 6Mb down/384Kb up. (Lowercase "b" for "bits")

Speed tests with 3 of the popular ones, all sites in San Francisco:
dslreports.com: (bandwidth provided by Megapath Networks)
8126 Kb down / 1582 Kb up

Speedtest.net (bandwidth provided by Unwired)
6823 Kb down / 781 Kb up

Speakeasy.net
8944 Kb down / 1195 Kb up

I also have my own Linux box at local University. I did a wget from both ends, watched the output. (No fancy TCP overhead calculations or anything)
For the first little bit I get:
8400 Kb down / 1544 Kb up
And then the Comcast traffic shaper kicks in and I get:
6176 Kb down / 356 Kb up

Which is more or less what Comcast advertises. For the most part, I'm a happy Comcast customer, although it's not like I haven't had my share of troubles. After complaining about a flaky connection a few times last year, they finally sent in a technician. He measured my signal quality, laughed and said, "I can't believe this works." Apparently the last cable guy Comcast sent over installed a splitter and cable that was never intended to carry a broadband signal.

This is actually a pretty dumb Ask Slashdot

[ben there...]

It explains it right in Comcast's literature. I also get 16+ Mbps on a 6 Mbps Comcast connection in the speed tests and while using newsgroups. It is designed so that you can download what I consider to be small files, like Firefox (5 MB), in a couple of seconds.

You can sometimes trick it into downloading at burst speed while you're already downloading by using multiple connections, although it doesn't always work. I still haven't figured out the exact criteria for it.

10 is right

[Anonymous Coward]

Comcast offers download speeds of 6 and 8 Mbps.
If you pay for 8, then they actually allow you up to 10 Mbps.
However, as soon as you go over 10 Mbps, they will bitch-slap you down to 8Mbps.
What this means:
- If you do many short downloads, you can get 10 Mbps. This helps remove congestion from the network.
- If you do one large download, then you are locked back to 8 Mbps. You cannot hog their resources.
- If you have a router with throttle control, then you can lock yourself to 10 Mbps and get much faster rates. But since most home firewalls cannot do this, this isn't a huge risk for Comcast.

Comcast's thresholding is managed on a per-TCP-connection basis. You can have one connection bitch-slapped to 8 Mbps and others running that raise you above 10 Mbps. I have measured as high as a cumulative 12 Mbps (11.7 something).

Comcast Also Lies

[oMaT]

Not to be a troll or anything but my experience with Comcast cable broadband has been eye opening as far as their business practices. If you order their straight-up normal 6Mb/768k connection it's $42.95 after whatever special they're running is over. After 4 weeks of daily testing of that connection not ONCE did the upstream speed surpass 384k. (And other /.'rs are right, their "powerboost" thingamahoober skews the downstream tests to the point of utter uselessness.) Using the Speakeasy test, the upstream "needle" would hit 384k and simply hover there. When I called and complained (multiple times, after being transferred around, hung up on, etc) I eventually hit a salesperson who, in essence, informed me that that is the best I could get without upgrading to their more expensive $59.99/mo package.
*Boop* *Beep*
...Aaaand lo and behold the upstream hits 768k with no effort at all. I now pay the greater rate and have for a while now and while I'm satisfied with the speeds and don't care all that much about the price, their bait-and-switch tactics are disgusting. So, anyone pulled into this thread who has Comcast cable broadband and is curious about their upstream speeds, I'd give em a call and raise hell. There *is* something they can do. While saturation may play a part in what you're seeing, there is plenty they can do as well.

Re:Comcast Also Lies

[cryptodan]

Sometimes the billing department doesnt receive the right product code, and thusly when you request the upgrade you get the wrong config. This happened to me once, I called and complained and was fixed 15 min later.

Comcast

[Mockylock]

Soon, Comcast will be offering 16Mb/s in all major cities, and gradually working to smaller towns within a few years.

When I lived in Reston, Va, my speed was 16Mb/s (2MB/s), continuous, without any initial boost.

It's being pushed in FIOS areas for the time being, only to compete directly with Verizon... but will soon be everywhere. (or so they say.) Within a couple months, the new area I live in is actually switching to 16Mb... but, I think they're waiting on FIOS as well.

First rule of usenet

[dj245]

is that you do not talk about afforementioned network!

Reston (Fairfax County), VA

[pvera]

Comcast, 16MB up, 2 MB down, is about $55 including the modem rental. This is a $10 add-on to the basic service, which they claim is 8 MB down, 1MB up. In reality it is about 13MB down, 1MB up. During busy hours it will drop as low as 6 MB. Anything under 6 MB usually means there is something wrong.

I have had Comcast since 2000, I have had maybe 4 real outages, none lasted a day. It is so stable that I freak out if it as much as loses sync.

Their only real problem is that until FIOS got here, they were the only game in town, so for example when the download caps nonsense started, we had no choice but to STFU. Now that there's competition, you don't hear a peep about either download caps, or about hosting from home.

Re:Reston (Fairfax County), VA

[Afrosheen]

"Comcast, 16MB up, 2 MB down, is about $55 including the modem rental."

Wow, that sounds like a fantastic service...if you are a bittorrent seed. 16MB up? 2 MB down?

I know, I know, you mistyped it, but that's a pretty glaring typo ;).

Re:Reston (Fairfax County), VA

[hedwards]

I have had Comcast since 2000, I have had maybe 4 real outages, none lasted a day. It is so stable that I freak out if it as much as loses sync.

Consider yourself lucky. Before we dropped Comcast as our broadband provider it was out daily for around 3 hours a day. That lasted nearly 6 weeks with barely even the smallest concern on their part. Basically they wanted us to call them each time for a credit. No hint at all as to when if ever they were going to fix the problems.

Cable service around here started to go downhill when Excite went out of business. The ATT service wasn't as good, but it was still good. Suspiciously enough almost immediately after Comcast took over things started to go downhill with our cable service in general. Right now I have Directv because Comcast doesn't even care about providing quality cable tv let alone any of the other junk services they provide.

It really depends where you are in a city, my brother who only lives 3 miles away has yet to have any problems with either his tv or his cable modem. There just isn't any incentive for Comcast to provide service as they don't have any competition in the cable market here. But I do see more and more dishes going up around here all the time. Especially in my neighborhood where the cable is crap and expensive and we have our choice of several DSL providers and a couple of satellite providers as well.

Re:Reston (Fairfax County), VA

[pvera]

My first weekend was like that, up and down nonstop. Why? Because the tech installed the splitter backwards.

I think another reason why it is so stable is that this county was one of the first national test beds for cable modems. They basically tore down most of the cable tv infrastructure for this county. I have no clue if this is how the rest of the country received cable internet, or if they simply plugged in new equipment and kept the old cables.

Re:Reston (Fairfax County), VA

[Mockylock]

I moved from Reston to Sterling not too long ago and Adelphia was switching to comcast. Unfortunately they didn't honor the 16mb download across the road. HOpefully they will soon. From what I've heard, Reston is about to get even faster, also.

Re:Reston (Fairfax County), VA

[Mockylock]

Yeah, Reston is known for being a bit more "high-tech" than other places (as well as Ashburn and parts of Herndon). It's not lucky to have good service out there, just a good location.

I urge caution

[u-235-sentinel]

Even if Comcast provides speeds of 50 Meg or faster I would be very cautious. You can download faster yes but what is the acceptable number of megs you can download in a month? Comcast has been terminating customers for years over this issue and it seems to be increasing.

Two people in my neighborhood have now been disconnected. Dozens more in the valley and hundreds I've spoken with either via phone or email. The company doesn't seem to know how to solve this problem. There has even been talk of a class action lawsuit. Should be interesting to see if that happens.

So far my blog has over 28,000 visitors in only a few months. If you have a story please post it there. We'd love to hear from you.

Comcast speeds, and sharing

[lundbergaj]

Comcast speeds are certainly pretty good when they're working right. However, with sharing, you also get problems if your neighbors have connection problems. For the past several months, my comcast connection has been down as much as it has been up. The very likely culprit is a cut cable line in my neighbors yard that was spliced, perhaps not very well. It generally seems to fail, both for him and me, most often when it's cool or raining. Even when it's up, it sometimes gets packet losses up to 40%. I've had the Comcast "techs" out 3 times now, but if the intermittent connection isn't bad at the moment, they can't find the problem. They don't seem to have the capability to find the problem by scanning the whole area for problems or doing any continuous scanning to fix the problem. I may need to switch to DSL shortly, as a slower connection is much better than an unreliable one.

Comcast HSI

[cryptodan]

I use Comcast in the Blatimore/DC Area, and I receive anywhere from 7 to 25Megabits for my download speeds depending on the server, and I also receive up to 2Megabits upload. Comcast has increased their speed to compete with Verizon FiOS. Speedboost is nice aint it?

wrong, wrong, wrong

[Anonymous Coward]

First, the person above that suggested switching cable modems is dead wrong. Cable Internet uses a standard called DOCSIS. DOCSIS version 1.0, 1.1, 1.1 Lite, 2.0 and 3.0 all have a UP Stream CAP and Down Stream CAP setting. In 1.1 BPI was introduced.

The rate increases people see when they do speed tests are the results of edge caching AKA Anonymous Proxys. So instead of going back to speakeasy.net for that large file, the local HFC (Cable Plant) has a cached copy of the same file. A simple URL query is done to make sure the file isn't out of date and if it's not, you get a cached version of the same file from a near by Caching server at the cable company's head-end.

twisted pair / coax

[Anonymous Coward]

Some of you pro-DSL people fail to realize that DSL is very limited. Twisted pair will always limit it. Verizon's FiOS and the cable companies will be offering up to 100Mbps downstream soon. There is no way that DSL can even try to compete with speeds like that.

One thing is for sure, I will remain with cable as it is VERY configurable, versatile and does not support the AT&T evil empire. I am so sick of the shady marketing campaigns of Verizon and AT&T. If people just did their research they would understand the following:

1. DSL service shares bandwidth amongst ALL users connected to the same DSLAM. Cable shares bandwidth amongst ALL users connected to the same CMTS.

2. DSL is limited to 18,000 wire-feet.

3. Even with vDSL (the fastest DSL connection BUT not available 90% of the time), you cannot get over 56Mbps with DSL.

High upload speeds for ~10 seconds

[GrandPoohBear]

I am a Comcast customer in Eugene OR, with a 6 Mb download service. I've actually gotten sustained downloads at near that speed, so I'm pretty happy, but in a lot of ways I'm more concerned about my uploads. When I first clocked my uploads, there seemed to be a hard cap at 384 kbps, but recently, speed tests showed closer to 1.5 Mbps (yay!). So I put it to the test last week by transferring some big files from home to work via AFP. Sure enough, for the first 10 - 15 seconds I was getting a transfer rate of upwards of 200 KBps. Then suddenly the speed dropped and there was that 384 Kb hard cap again. So I'm wondering: is there a mad conspiracy to allow large bursts of traffic for long enough for the speed meter to complete, and then cap it after?

comcast speeds

[cetae]

I pay for the 8Mbit/768Kbit service from comcast, and on multiple occasions I have gotten OVER 1MByte/sec while download a torrent with uTorrent, I cap my download limit to 800kbytes/sec now otherwise utorrent will completely saturate the line and I can't do much more.

I believe it is possible that those speedtest ratings are correct.

I'm in the northeast...

[SuperBanana]

As someone who used to work for Comcast allow me to say rotflmao. Either you are one lucky sob or you are lying. Just as a matter of example (one among many) during the entire nine months I worked for Comcast the entire state of Illinois never left the outage board.

I don't live in Illinois. I and my parents live in the northeast, which is where MediaOne started. That's probably why we've always had decent service. That and the fact that Comcast answers to the public utility commission state-wide, and in Boston, the mayor's office.

We've certainly had some outages, but they were brief (few hours) and half the time, weather related (ie, major storm.) We've even received partial service refunds if we complained enough.

Lets be a little more smart about this

[bonez_net11]

I recently downloaded a 80GB laptop backup image from a server at work (who has massive amounts of bandwidth) to my home Comcast... I downloaded via sftp2 and the final speed was 2.0MB/sec. Yes, two megabytes per second. This mostly copied overnight. I -do- pay for the "8mbit upgrade" and this was a continuous file (single connection to port 22 of host machine) transfer. The bandwidth limitations seem to be 'soft' whereas if the network is not busy it gives you more. I have a 10ms ping to said machine at work and have had other files (on different days) download only at 1.2MB/sec. So, my experience is somewhere between 12-20 megabit per second on my Comcast. I'm in Lynnwood, WA (just north of Seattle).

Why you get fast speeds...

[dudeX]

It's easy. If Verizon FIOS is readily available in a neighborhood with competing cable, magically your speeds go up.

When NYC proper gets FIOS city wide, I expect my Time Warner Roadrunner service to automagically give me at least 20mbit/5 mbit speed.

Until then, I'll just have to be content with 10mbit/768kbit.

My brother lives in Westchester county, NY and he gets 10/5 from Cablevision, because FIOS is readily available in his neighborhood.

Normal behavior

[DerekLyons]

From TFA:
  Here are the reported download speeds in Kbps: New York, 18,946; Washington, 15,821; Atlanta, 11,257; Chicago, 10,042; San Francisco, 4,230. What is going on?
 
I see pretty much normal 'net behavior. Different sites have different download speeds because there are different paths between your computer and the server. Duh. (And if you repeat the test - you'll almost always find that the results vary between tests as internet 'weather' conditions vary from moment to moment.)

Logical Fallacy

[Richthofen80]

Just because A precedes B, does not necessarily mean that A caused B.

Re:Logical Fallacy

[WindBourne]

no, but it is an implication, not a tautologoy. Or do we wish to argue semantics?

Hate Comcast but their speeds are nice

[Serpent Mage]

I don't do test sites as they are meaningless to me but doing bittorrent based downloads a lot, and being able to download at peak rates of 2 megs per second (usually around 1.3 megs per second) I gotta say it is pretty good. I can download huge files (gigs range) in a matter of minutes.

"Reliable"

[fm6]

THE INTERNET . It's unreliable, not guaranteed, never has been, and YOU ALL KNOW THIS AND HAVE BEEN TOLD IT OVER AND OVER AND OVER.

I seem to have missed that meeting.

Anyway, bullshit. Every job I've had for the last 10 years wouldn't have existed if the Internet weren't reliable. Not 100% reliable — no technology is. But if the people who manage your network connection and the remote systems you access are at all competent, your Internet connection is going to be as reliable as a well-run telephone system.

Or maybe you mean that Internet technology doesn't have a built-in guarantee of bandwidth. True, but trivial. No communication system can guarantee that you'll get through if it's being used beyond its capacity. The phone system can't guarantee that you'll get through if everybody decides to use it at once. But once again, if both your Internet provider and remote system provider know what they're doing and have anticipated demand for their services, you can count on the system working.

I know: you're referring to the fact that most low-level networking technologies do not guarantee that any given data packet will get through. This is "unreliable" in a technical sense, which has nothing to do with "reliable" in the human sense, any more than you can use a synchronous clock to tell time. In networking "unreliable" just refers to a network design philosophy that says that it's more efficient to resend packets that didn't get through than to guarantee that all packets will get through. That's sort of counterintuitive, which is why it took so long for "unreliable" Ethernet to displace "reliable" Token Ring. But it did, and now a good chunk of humanity relies on Ethernet, even though it's "unreliable".

Re:Time of day? Bad analogy!

[Anonymous Coward]

It not a bus (truck), it like a subway (tubes)!

Built into the system

[IBitOBear]

Actaully, since the lower performance package limits the upload speed by a greater percentage(*) than the download speed, and since the upload speed can be shown to be the dominant factor, they don't have to do a single thing _technically_ to make you "Feel The Pain".

Between that, and the (probable) QOS tiering differences between the two packages, you will also Feel The Pain naturally as your cable modem segment comes under increased load.

They almost certainly want to get the data out of their systems and into yours as fast as possible so that they can have cheaper systems on their end. Backlogging your data behind a throttled link takes memory. It actually costs them money. It's better to set up the systems so that the load characteristics meet the promised numbers in demonstrable but very raw/low-level way and then just move the data as fast as possible. That way if someone tries to sue them about performance they can go "look here, we set it up so that at peak load everybody gets their share, and so what if they _could_ get more when things are idle, that's not bad for the customer". Meanwhile their actual load conditions end up being limited by the much more practical, distributed upstream throttles in the individual modem initialization settings.

This is further demonstrated by the incidents where hackers hacked on their modems and got much higher levels of service. If the systems at the distribution point were actually throttled, tweaking their modems would have only amounted to a slight increase in total throughput. They'd have had to hack the central distribution facility systems as well. (Which they did not have to do.)

Disclaimer: This is a black-box analysis on my part. e.g. this is a well-educated guess based on observations. They may well be doing something stupid with their central facilities. I would only cost them a lot of money for no real benefit (they get charged for wasted retransmits at _their_ meetered access to the internet at large too, so dropping data at the boundaries of the cable segments would be damn stupid). Then again, they have to be wasting all the green they make for overcharging me _somewhere_... 8-)

(*) Consider that when the download speed goes from 6 to 8 Mbps (a 33% increase) the upload speeds goes from 384 to 768 kpbs (a 100% increase).

Speedtest sites not so speedy

[Anonymous Coward]

The problem with *almost all* speedtest sites is that they only test one connection at a time. I am located in Europe, and downloading from SUN.COM East Coast gives me like 12 mbit. Starting other downloads (for the other parts of Opensolaris DVD), it hits the ceiling at 100mbit-Ethernet's physical limit near 75-80 mbit/s.
The speakeasy-powered speed test site speedtest.net gives me boring values in the 3200kbps area for download for the East Cost. (Tested New York and Raleigh.)

Crap upload

[Ramjet350]

The problem with these broadband companies is the upload. How useful is 19Mbps really when most sites can't even support it? It sounds great on paper but I would really rather have more upload. 384K that Comcast offers in Denver is a joke.

20 Megs during in-house tech call

[TheRealJobe]

I experienced a similar feat with Comcast in the KC area. I moved from an Apt about 4 miles down the road to my first home. In the apt would see an avg of 600k down. When I moved out to the house I noticed that everything seemed significantly slower... So I ran out to speakeasy and avg'd about 4 down and 756 up. This combined with the fact that my TV's on Demand always timed out was enough to have me ask that an in-house tech come over for support. The in-house tech was to come over Saturday afternoon, so that morning I made sure to re-test the line. The results had been similar at speakeasy, 4down and 756 up. This is where the "Magic" begins: The tech shows up, sits in my driveway for about 10min on his phone, comes directly inside and runs a diagnostic video on my OnDemand service. Said there's no problem and it must be a peak usage bottle-neck. ....okay peak usage all week? Which is still intermittently occurring to this day... We then go to my computer and fire up speakeasy *pixie dust* I suddenly increase speed 5 fold, score 20 down, but only 756 up; I tried all the speakeasy test sites. Most range from 20-16 down with 756 - 1.5 up. With the exception of Seattle with shows 5.5 down, and 756 up (probably real speed.) So I question the installer why... he too said it's the "magic" of speed boost. Post his departure I still rock the house at speakeasy but lag everywhere else... I guess the real "magic" of speed-boost is it only works on equally magical sites :( I have no doubt in my mind that they are treating the speed test traffic preferentially.

Especially...

[Kamokazi]

Especially when you figure in *AA legal fees and settlements.

Kernel.org

[moosesocks]

My favorite "quick and dirty" bandwidth test has always been to download a full bzipped linux kernel from

The link's always right there on the homepage, it's uncompressable, it's a pretty big file, and Kernel.org itself has more than enough bandwidth to saturate any connection you'd typically find at home.

You also won't catch the ISPs trying to cache it.

Of course, if you want a 'real' test, Speakeasy's great, although I'd be weary of an underhanded ISP caching the speed test.

Comcast speed

[Averyge Joe]

I have several clients with Comcast (various flavors) and am absolutely amazed at the download speeds. I haven't tried them on any speed tests but I can say confidently that I do not fear 150mb to 250mb downloads...

Very strange...

[Anonymous Coward]

I'll never understand why so many people pick on Comcast. I think they are wonderful. Where I live it's either Comcast, DSL, or Satellite and if you're not on a main road or close to town the odds are you can't get Comcast or DSL and the satellite companies are booked solid so good luck scheduling an installation. I consider myself extremely fortunate to have Comcast, the download speeds have always been satisfactory, it's very rare that there's any downtime, and they give me Mcafee virus protection for all of my computers at no charge. On average, I get consistent 500kbps download speeds, I have seen sustained downloads as high as 890kbps using bittorrent, and I have seen momentary spikes of between 6,000-10,000kbps. Comcast is A LOT faster than DSL and the DSL interferes with the telephones, which they say it isn't supposed to do BUT IT DOES. Satellite has decent download rates but very high latency and it's affected by weather.

Yeah Comcast doesn't give you a fiber backbone but for less than $100 a month their service is pretty damn good.

Comcast Sucks

[neuromancer2701]

I live in Alexandria,Va and we are blackmailed into using Comcast because there are no other providers for cable or broadband internet. I have had nothing but problems with setup and customer service, but at least I have good bandwidth. I average about 400KBps but can get up to 600400KBps with BT. The one thing that needs to be changed is the whole TV licensing municipality. Comcast and Cox have this unsaid agreement don't step on my toes I won't step on yours and Verizon does not want to apply for a TV license in all the municipalities. If I could get Fios I would drop Comcast at the drop of a hat.

Comcast tests

[Anonymous Coward]

Here in San Jose I pay the extra for 8Mb (over the standard 6Mb). What I actually get is 10Mb. With the "PowerBoost" I see initial peaks past 45Mb.

What I typically see when I start a download from usenet is it skyrockets up to 45 or so, then arcs back down and settles at 10Mb, flat, all day long. The PowerBoost is meant for short file transfers, and for this it works well, but once it detects that you are doing a long download, they switch you to your actual maximum rate.

With the Speakeasy test I regularly get 25Mb, and with the java app mentioned above (anl.gov), to Santa Cruz, I get about the same:

running 10s outbound test (client-to-server [C2S]) . . . . . 1.58Mb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . 23.09Mb/s

The speed tests are inaccurate (too high) because they are too short. They always test in the PowerBoost timeframe. So yes, maybe part of the reason for the PowerBoost is to make the bandwidth tests look good.

Also, a few months back Comcast quietly upped my upstream from 768kb to 1.5Mb. I know Comcast is supposed to be evil, but right now I can't complain. I have Ethernet rate downstream, and T1 rate upstream.

Proper speed tests

[jarome]

The PC Magazine tests were really bad. First of all, tcp/ip cannot get up to speed on small files. If you want to fill a really big pipe (say 10 Gbps), it can take an hour to get up to speed!. Second of all, the route to the chosen Web servers was different depending upon where the user was located. A proper speed test that will actually help diagnose things is at
http://whisper.cs.utk.edu:8234/

With Comcast I get (at the moment) I get 2416 Kbps outbound and 15315 Kbps inbound. This will drop after about 20 seconds of transmission due to the speed boost which has recently been put into effect on both inbound and outbound connections.

So PC Magazine has unfairly smeared Comcast, which is now giving stellar service.

Another issue is that Comcast uses AT&T's network, and when AT&T bought BellSouth, they agreed to net neutrality, which means that they now peer at many more locations with other networks. It used to be that my packets went twice across the country to go 10 miles from home to work. No longer :-)