Blog Post

Need For Speed… How Real?

Stay on Top of Emerging Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

After years of being stuck in the slow lane, the US consumers are finally going to get a massive speed upgrade and taste the true broadband for the first time. From a 512 Kbps world to 6 Mbps, then 8 and soon 15 Mbps…. it seems the future has finally arrived. And with that, the question…. how much speed is enough? Can we the consumers really tell the difference between 15 and 30 Mbps? Or is it just a way for the broadband operators to get us to pay more… for something which we might use less.

To say that we are a nation starved of bandwidth would be an understatement. Generically speaking, The average US broadband experience is stuck somewhere between 500 Kbps to 3 Mbps. However, as we have been reporting in the recent months, things might be changing, as cable operators, and phone companies both roll out faster broadband connections.

BellSouth is now selling a 6 Mbps while Verizon is offering a $179 a month 30 Mbps plan. Comcast customers can now dream of between 6-and-8 Mbps speeds, while Cablevision has offered 50 Mbps service for an undisclosed amount of money.

The Wall Street Journal has a nice round-up of some of the speed upgrades. But what is behind the new found “speed thrills” philosophy of the incumbent carriers? One word: money. It costs the network operators a tiny bit more to offer more bandwidth, but they can sell higher speed connections for a premium price. Actually, the more speed you give to the consumers, the less they use the network. (Of course, no one can really tell if you are getting a real 3 or 6 or 10 megabit throughput. Support people at broadband providers have a standard line: its de innernet, what can I do?)

David Card, analyst with Jupiter Research points out that, “Time spent online isn’t shrinking, nor, in most cases, is the absolute number of users of any given function going down. It’s just that the breadth of activities any one user engages in is shrinking.” That perhaps can be explained by what Dave Burstein, publisher of DSL Prime wrote in an email to me …

Websurfing runs at only about a megabit per second, and nearly everything else except downloading is effectively throttled down at the source. Downloading turns out to have some natural limits as well; at 100 Mbps, you can download enough music for 24 hours of listening in only four minutes per day. The practical result, confirmed by high speed leaders like Masayoshi Son of Yahoo BB in Japan, is that the faster speeds yield only a extremely modest increase in real traffic demand.

As part of my daily reading, I read a fantastic essay by Shing Yin of Bernstein Research. Is Broadband Speed Like Money? Great line…. isn’t it?

….are we about to build ourselves into another bandwidth glut? Not in the backbone…but in the access network, where the Bells and cable MSOs are racing to roll out ever-faster broadband speeds? ….

Extending Shing’s methodology , I put together this little chart.

What this shows is that as we increase the speed, the real impact of the speed on what we do with it is marginal. Can your eyes tell the difference between a web-page loading in one second or 0.27 seconds. I guess not. If you can download a music file in 1.08 seconds, does that really mean you will be buying music all the time. No you perhaps will be buying better quality, and perhaps marginally more music. There is the other option, but its just easier to pay! Sure at 30 Mbps you can download DVD quality The Bourne Identity in 11 minutes, but its still going to take you 2 hours to watch it. These are analog questions in an increasingly digital world.

“Can a consumer tell any difference between 1.5 megabit per second and four megabit per second service?” asks Bruce Leichtman, president of Leichtman Research Group, Inc. “The answer is no.

So what the incumbnets are selling is a perception of speed that thrills. Don’t get me wrong…. I will upgrade, and hope the experience improves, but at some point, we need the applications that truly harness this speed come-along and are allowed to thrive. Not likely in the “we will control the net” attitude adopted by the incumbents. Even in truly immersive multiplayer games, its the latency, not the speed that matters.

The real bandwidth question is when are going to see an increase in the uplink speeds? Since the incumbents throttle the uplink speeds to barely usable, the broadband remains quasi one-way. P2P, the one true broadband technology is yet to blossom, especially in the legal realm.

Meanwhile perception is reality…. so more bandwidth.

61 Responses to “Need For Speed… How Real?”

  1. I think you might be looking at this the wrong way. Perhaps content is being held down by bandwidth. Eventually, as the bandwidth increases the offerings online could grow in size to match it.

  2. My university has fiber-optic gigabit. I can get a 700MB video file from someone on campus in, oh, a second or two. I think 7200 RPM hard drives are starting to be a speed-limiting factor at that point. The I2 Hub (internet2), when it existed, had similar results for transfers across the nation. It’s OMGwonderful when you can do things like remote desktop / FTP / grab linux ISOs nearly instantly with no snags, as I’ve found myself doing when i have 2 minutes before I have to leave for class. For people like my parents who check email and surf the web, a 3 MBPS cable connection may be fine for a while longer; Serious computer users, though, should not be happy until enormous transfers (HD video) take less than a second and game latencies are all below 100 ms.

  3. wads, i am not disagreeing with you or anyone. the point i am making: downlink speeds don’t mean anything because in the end uplink is the choke point. secondly, when video comes, the bandwidth will be needed. right now the 15 megs stuff is all just a way to get more money from you pocket.

  4. I think you are missing something here – you are thinking only about casual use – surfing – even downloading. I want more speed for work. Working at home will only become a reality when bandwidth allows communication both up and down at very high speeds. I want ethernet 10/100/1000 speeds to communicate with the server at work.

  5. Yeah, and as Bill Gates said in the 80’s, no one will ever need more than 640KBytes of RAM….

    You are suffereing from the innovator’s dillema. More Bandwidth is like more Memory. You can do more innovative things with 10s or 100’s of Mbps that you can’t even imagine when you are stuck in the Mbps range.

    So the countries that are stuck in the 1Mbps range are stuck in the non-innovation zone.

    And I don’t buy that the RBOCs and the CableCo oligopoly is going to get us higher speed broadband any time soon.

    I still can not get ANY cable modem or DSL service in Saratoga California, right here in Heart of Silicon Valley.

  6. As Martin Geddes ( recently said: Demand easily exceeds realtime, if it is affordable. ” Darling, we are leaving in 15 minutes, did you download these 5 movies for the kids in the car? “”Oops, will start it now”.
    And for anyone who has looked at the 3D-displays-wthout-funny-glasses coming up (Fraunhofer institute, Philips e.a.): 100-fold increase in demand of bandwidth compared with HD? Imagine that two-way…after all we are animals who are dying to communicate, anytime, anywhere.
    The garbage belt of history is densely populated with anyone who proclaimed “There is no need for anything more/faster”. Usually thats a sign that one is growing old…..

  7. Some folks are streaming HDTV in the house and this is why we need more bandwidth to the home as well as Ethernet over Coax in the home. We stream HDTV via WiFi today; last week we reported how it worked on 802.11g and today we reported how it worked on 802.11a. It can be done, but more pipe is what we want!

  8. Well, I’ll agree that, while surfing the web, one can’t really tell the difference between old fashioned ADSL and ADSL2+… But having spent a year in France with an ADSL2+ connection, getting about 15Mbps (compared to the 512kbps I was getting in Australia – and soon have to return to!) I see a big difference, but in different areas.

    While I’m surfing the web, sure I get a slight increase, but it’s not huge – certainly not enough to pay extra for… And when I download things, I get faster downloads (though only really when I’m downloading off a local – ie French – mirror site), but because I’m not downloading huge amounts all the time it’s not a huge issue

    I see two areas where bandwidth will make a significant difference:

    1. connection sharing: I’m the only one using my 15meg connection, so it’s rarely running at its potential, but if I was sharing it with family we could easily have a couple downloads, a couple background email/IM, and a couple surfing the web all at once…

    2. television (and other services): with my connection, I get seamless television and VOIP bundled (by seamless I mean: aside from plugging into the ISP-supplied box, an average user wouldn’t know the technology behind their tv and phone); I can watch tv while downloading and surfing the web, without hassle. My ISP has also just released a video-on-demand feature, though I’ve yet to test its performance…

    What does increased Speed give? Opportunity. Average users may not know the technology, but they will be increasingly attracted to connectivity packages that bundle tv and phone options…


  9. Some thoughts:

    1. Users can detect 1/10 second delay as meaningful latency.
    2. 640K turned out not to be enough.
    3. Availability, like anything else in life, eventually changes supply and demand to the extent that new kind of applications become both possible and desriable.
    4. Not too long ago the around the web was how people were restricting their RSS feeds or even thinking of abandoning them because of bandwith problems.
    5. To this day video/3D/info streaming remains a major bandwith problem.
    6. When setting up capacity/throughput plans at the serving side, we make assumptions about current, limited available bandwith.
    7. The functionality and efficacy of P2P networks are critically impacted by limited bandwith.

  10. I want near-infinite SYMMETRICAL bandwidth infinitely cheap. And I want it now.

    If the webOS/web 1.1 (oh, sorry, web 2.0) crowd wants the whole webapp thing to materialise then they want that too.

    And the download/P2P and especially bittorrent crowd will want it too.

    And corporations (offsite backup, no more mail limits, the list goes on).

    And anyone who works with large amounts of data will want it too.

    But the speeds you mention at the prices you quote: BIG YAWN (says the EuroSnob :P)

  11. I just wanted to point out that ISPs are (finally) starting to increase upload speeds as well. The Tampa Bay region of Brighthouse has 2 new packages that are only available with bundling services. 10 down/1 up and 15 down/2 up for residential services.

    Also commercial class is seeing (unadvertised and unannouced but active at this point) speeds up to 15/2 dependin on what packages they currently have.

  12. Joel Johnson

    Wouldn’t you know the last comment would say what I was excited that no one else had said (so I could look smart).

    We’ll use as much bandwidth as can be thrown at us, without a doubt. As soon as the majority of American households have 30Mbps connections, it’ll be time for the next bump in video resolution—up to the 4K stuff, probably.

    Then those torrents will be just that much bigger.

  13. Wow, I remember when people used to say, “Why would I ever need more than 640KB of RAM?” or “Why would I ever need a hard-drive larger than 50MB?”

    Somehow though, we figured out a way to not only take advantage of the increased capacity, but I would argue that our computing experience is even better today than it was back in 1980… :)

    Seriously though, it’s the same thing with bandwidth; technology will scale to ensure that the bandwidth we have is well-utilized. And it’s not a stretch to conceive of how that might happen.

    With streaming video, it’s easy to imagine bandwidth needs in the tens of Mbps; in fact, if you want to stream production-quality high-definition video, you’d need on the order of about 20Mbps downstream…and this is just a single stream. Sure, you can argue that nobody is currently streaming HD video, but it’s a chicken and egg problem, and I promise you that the content/technology won’t be generated until the bandwidth is available.

    And, of course, as other have pointed out, the upstream speeds will need to improve greatly before the downstream speeds become nearly as important.

    This is a surprising argument coming from you, Om. Were you just looking for something to write about today? :)

  14. Yea, but… sub-second responses will be important for developing web applications. A second for a page to load (or for a response from your XMLHTTPRequest) sounds like nothing, but can seriously interupt the flow of getting work done.

  15. A very thorough study done by McDonnell Douglas Health Systems in the ’80s showed significant improvements in individual of programmer productivity when terminal service response was decreased to sub-second levels.

    While this study was done in a mainframe environment, I’d suggest that the effectiveness of sufficient bandwidth to the end user has similar value.

    Obviously, the limitations imposed by the SERVER/CLOUD side of the equation will be exposed when the pipe isn’t the throttle. But, assuming web services providers (not to be confused with the network ISP) keep their bandwidth and server resources consistent with the desired consumer experience, my satisfaction will be greatly improved by larger pipes. And my productivity. Even envisioning the effect on the simplistic scenario of acquiring the hybrid Google maps image I want this moment stirs my glee. And it gives rise to optimism that Web2.0 (and beyond) services will attain real value.

    Though I’m not so naive as to think that milking the cow isn’t part of the equation, I still believe capitalist market forces will drive speed/$ models down for the consumer. Especially since services like multiple streams of demand TV and uploads of video telcom. activities etc. will drive the web services providers to demand that their consumers be provided the necessary quality of service.

    If the Bells, etc. don’t respond to their major clients and provide an effective cost model for the ultimate consumer, they’ll probably be bought/brushed aside by the services vendors themselves.

    I agree with McNealy, the network is the computer. It just hasn’t pushed past the IBM/386 equivalent, yet.

  16. I agree that the difference is NOT at the web (page) level.

    From my experience here in France, where most of the DSL offers are now at around 20 Mbps (ATM), the main difference is for video. There are more and more video contents (I mean legal ones. Specially video blogs, trailers. Not mentioning IPTV and VOD) and having a fast download makes it a totally different experience. You won’t spend much time on video if you don’t have a very fast connection.

    Also keep in mind that most of these offers will come with TV soon. It takes about 3-4 Mbps for one tv channel feed. If you want to watch different channels on different TV, here we go, you are already running out of bandwidth :-)

  17. While I have my doubts about the need to “stream” anything, I think Golfingguy makes a good point. What happens when I have five HDTVs watching five HD programs simultaneously?

    Anybody see the HANA annoucement last week?

  18. I am shocked this question comes from GigaOm! Will you promise that you will never upgrade beyond 6 MBPS? Me, I won’t be satisfied until everything is instant (sadly, I admit – hopefully I will get more patient as I get older) but more importantly I bleieve speed is the enabler for innovation we have not yet imagined.

  19. We need to continue increasing speeds until multiple HD programs can be viewed “real-time” in the home.

    Lets not get short sighted by looking at current behavior and forget we want complete convergence – we’ve been asking about it and talking about it for a while now. Lets stay consistant.

  20. I thought many ISPs in the US already offered a pretty cheap broadband connection? Probably I’m wrong on that.

    But to think about it, this faster connection is the next business enabler for the net. It drives a significant growth on what can be done via the internet and it’s exciting!

    I just hope that we won’t face another burst ;-)

  21. For me, one of the big issues isn’t the download speed, it’s the upload speed.

    I recently moved from Comcast’s “6Mbps” service which often had its upstream bandwidth capped at 128Kbps (and seldom was much better even when not capped) to Verizon FIOS, with 2Mbps of upload on a 15Mbps service.

    As a developer who’s often stuck waiting for slow FTP uploads, this made a big difference.

    Of course, it’s not just people like me that this matters to — a lot of common consumer protocols (BitTorrent comes to mind) are often limited far more by the upstream bandwidth on an asymmetrical than they are by the download speed.

  22. Thomas Buccelli

    It seems that for the most part, download speeds are mostly for bragging rights between the various parties. With the exception of downloading various Microsoft service packs where you definitely can tell the difference between 6Mbps and 3Mbps, either are fine for general net usage. The big difference is with upload speeds.

    Verizon is offering services with 2Mbps upload speeds, while SBC and Comcast in my area top out at 600Kbps and 768Kbps with their high-end packages. More reasonably priced ones are at 384Kbps. I try and work from home when I can and this slow upload speed is noticeable. I am not trying to run a server of any type, but just connect to corporate network and transfer a few large documents or emails.

  23. I think your point is bang on. Higher speeds, beyond a certain point, are simply a marketing tool — especially when cable operators and telcos do everything they can to make it hard for you to use all that bandwidth, such as using “traffic shaping,” preventing you from running a server, etc. They’d like you to pay for it, but they really don’t want you to use it :-)

  24. Nice analysis, Om.

    I agree that the true value in these types of speed increases would only truly be realized through a disruptive innovation created to take advantage of the new levels of service. Anything that changes the scene so dramatically is begging for a creative mind to dream up a new way to take advantage of the network capabilities. However, it begs to question why consumers would pay for such speeds in the interim. I also agree that the chokehold on upload speeds is a huge limitation that needs to be remedied.

  25. Kashif Haq

    i agree that ordinary (non-techy) can’t tell the difference in downstream speeds….try uploading a picture or a file and one will quickly find out the difference in speeds/user expereince…glut in the downstream doesn’t translate into glut in the upstream :-)