Back in the late 1990s I was often asked what I thought would happen if Internet bandwidth was infinite — what would that change about the Internet itself? Level 3’s (LVLT) recent decision to slash prices on its content distribution network and rumors of new multi-terabit […]

Back in the late 1990s I was often asked what I thought would happen if Internet bandwidth was infinite — what would that change about the Internet itself? Level 3’s (LVLT) recent decision to slash prices on its content distribution network and rumors of new multi-terabit cables across the Pacific have me wondering if we are actually getting closer to having infinite bandwidth.

But when replying to the infinite bandwidth question I was prone to posing a return question — what does infinite bandwidth actually mean? As an example, I would often say that infinite bandwidth meant that I could personally consume terabits of bandwidth at any location on the planet at any time without loss or jitter. And, since bandwidth would be infinite, it would be the ultimate commodity and free.

I have often found that people who talk about infinite bandwidth lack a basic understanding of what bandwidth is — many think that if you have infinite bandwidth you can have instantaneous access to every computer around the globe. People often forget that the speed of light is very fast (around 186,000 miles per second in a vacuum) but nevertheless a constant. As an example, if you have infinite bandwidth between two points it’s like having a freeway with infinite lanes but your car can only go a finite speed. There is no congestion or traffic jam, but it still takes time to travel from source to destination.

On the Internet, a bit of data does travel quickly over fiber optics, but there are many different mediums and pieces of equipment between any given source and destination to slow things down. For example, from here in San Francisco it takes a bit of data around 320 milliseconds to travel to Bangalore and back, and about half that for a bit to make a round trip between here and London. Of course, on today’s finite Internet, these travel times vary with the time of day, the exact path taken and a thousand other networking variables, but I believe that these latency times are in the ballpark for most networks. Infinite bandwidth would not have this variability but would still be hampered by the pesky issue of speed of light between any two points on the planet.

Other than the speed of light issue, when people talk about infinite bandwidth they contend that accomplishing this on the backbone of the Internet (across network providers and between exchange points) is definitely feasible, but true end-to-end infinite bandwidth is hard to imagine given the constraints of the last mile in many locations. While I agree that the last mile is clearly a current constraint to infinite bandwidth, I have some hope that some high-speed variant of Ethernet or a wireless technology will solve this problem in the not-too-distant future — just look at what’s happening in Tokyo and Seoul.

So, in my mind infinite bandwidth is possible. Some thoughts about what we could do with it include: unlimited High Definition two-way video streaming (not like little YouTube screens, but to video walls), immersive and lifelike collaboration environments (think Second Life on steroids), limitless file transfer and backup (no more burning data to DVDs), real-time photo sharing (grandmothers will love this), and complete data mobility for both personal and work information (regardless of what type of data you have, you can access it from anywhere).

As the saying goes, “Necessity is the mother of invention.” So we need to know that bandwidth will be infinite and we need ways to necessitate this technology soon. How do you think infinite bandwidth would change your behavior on the Internet, either personally or for your company?

Allan Leinwand is a venture partner with Panorama Capital and founder of Vyatta. He was also the CTO of Digital Island.

  1. It would certainly be nice if the ISPs would stop complaining about those of us who like to download 500 gigabytes of data every month.

    1. This commonly occurs for apostrophes and quotation marks when they are written with software that uses its own proprietary non-standard code for these characters, such as Microsoft’s Smart Quotes. ,

  2. What is happening in Seoul?

    I’ve lived here for 3 months now and think it is totally over-rated. Unless you speak Korean, there are no useful webservices for English-speaking expats.

    Besides that, lifestyles are no different than any large western city. I’m from Dallas and there were just as many people attached to their cell phones and media players.

    The same goes for PC Bangs, unless you speak Korean, it is a major disappointment. Besides do you really want to sit in a smoke-filled noisy room filled with unemployed starcraft players? It’s not nearly as romantic as Wired magazine paints it.

    Here’s a piece I put together detailing the economics of this situation: http://www.mises.org/story/2661

  3. Infinite bandwidth would be infinitely expensive to build unless every element was free to build. Did I miss Cisco’s merger with Oxfam…?

    While a number of elements are considered here, the article does not cover the whole distribution chain (note link is uk specific).

    Consider for example the electricity required to power an infinite amount of bandwidth and serve an inifinte amount of content. Do we have inifinite energy reserves? Nope, sorry.

    Bandwidth has never been free, and it never will be. Perhaps 14.4kbps is “free” (to a number of significant figures) but just as the price of that declines, so the the bar gets raised… The net result is that you pay about the same for an ever increasing amount of capacity. You will never get an inifinite amount unless you get infinitely old… :-)

  4. Wow !! I just learned that the light has a speed limit.

    What’s the point of sensational headlines without content? Do you think it increases circulation? I would say you will loose some.

  5. Tim – in Seoul I was referring to the technology advances on WiFi and WiMax. From what I’ve seen in Seoul the government and various companies are working on some interesting last mile technologies. That being said, I am well aware that the end result does not match the hype yet….

    Jeremy – I’m not sure I agree that only getting old is the answer :) If an enterprise can get a 1Gbps of capacity from A to B for $10/month is that close enough to free? Granted, it’s not $0, but I think we’re getting reasonably close in the coming years.

    Satish – I was not trying to be sensational or content-free. Rather, I was trying to provoke a discussion on the uses of infinite (or nearly infinite) bandwidth. Any thoughts what you would do with infinite bandwidth?

  6. yes, extrememly boring post…. i would have expected more from this site…

  7. Infinite bandwidth would be nice, for sure, but it is a dream, and it will remain a dream for years.

    Here in France, and generally speaking in Western Europe, we benefit of very interesting broadband offers : Living in Lyon, I have an ADSL access with 16 Mb/s downstream + 2 Mb/s upstream + unlimited volume of data exchanged for not much than 15 € per month (OK, about 20 $ because the € vs $ is higher than ever. And within less than a year, thanks to FTTH deployment, I’ll have not less than 100 Mb/s downstream AND upstream, unlimited volume, for 70 € per month, about 100 bucks :-)

    BUT ! when I want to buy and download movies from VOD platforms, I still have to pay at least 4 € per film. It is twice the rental price of a DVD at the shop around the corner, despite the fact that I pay for the PC and the broadband access and the VOD platform does not have to pay for the shop, the DVD stock and the guy behind the desk. And as you imagine, the emergence of HD movies is a real issue for VOD platforms…

    WHY is VOD so expensive?

    BECAUSE of bandwidth cost!

    A company operating a VOD platform, serving a huge amount of terabytes per month, either through regular download, either through streaming, has to pay a very expensive fee to the telco providing the Internet connectivity to its VOD servers. The bandwidth costs a lot because telcos have to increase their network capacity as much as they increase the bandwidth of end-user access, for millions of households : they buy tons of core fiber network equipments from Cisco, Alcatel or Huawei, and they will continue to have to invest a lot within the next few years…

    The cost of the bandwidth and network capacity for content providers is one of the reasons explaining why our company is working on a new system, allowing to distribute movies as “nomad files”, circulating freely on IP networks, using every protocol (HTTP, FTP and more interesting P2P) to dramatically reduce the distribution cost of digital movies…

    The other reason is to provide a smart answer to this question: “How to allow Internet consumers to easily find interesting contents (movies, music etc), to freely discover a movie or a music album with a preview, without any piracy risk for the right holder?”

    DRM technologies are a huge mistake: they lock the content on a device, whereas the Internet is a way to share and recommend the content from a consumer to another. Why not let the consumer be a part of the distribution channel?

    Some solutions are emerging to replace DRM by “DUM” – “Digital Usage Management”: files can be freely copied, shared, redistributed among users, but their use remains under control…

  8. Hi Allan – $10 per Gbps might be ok for today’s applications, but the point is that by the time most people are using 1 Gbps because they can afford the $10, there will be some requirements that need Pbps, like the CERN particle accelerator. Even at $10 per Gbps, that’s still $10m per month for one of those.

    It will never be free because price only comes down as demand goes up (not the other way around) ;-)

    It may be that there is infinite demand – machines will use everything they can, even if there is a point at which people can’t. Mike O’Dell (Chief Scientist at UUNET way back) referred to the machines as cockroaches – he was bang on. P2P and collaborative computing are two examples today…

    Supplying this demand (whatever it is) means glass and it means silicon and it means power. All of which costs money, so the limits, once the infrastructure is in place, are commercial.

    Lets say the low end pays $10 for 1 Mbps today and the high end pays $10,000 per Gbps; when you get to the point where John Smith is paying $10 for 1 Gbps home broadband, XYZ Co is still going to be paying $10,000 but now they will be getting 1 terabit per second.

    I agree that much of it is artifically constrained – do you sell 1 Mbps for $10 or 1 Gbps for the same price? The trick for telcos is to price it so that people don’t think about downgrading – you need to keep Mr. Smith paying his $10 per month and XYZ co their $10k. You don’t want them to think, “I’ll just get 100 Mbps for $1 per month, I don’t meed more than that…”

  9. Jeremy – I agree that today machines greedily consume all of the bandwidth that is deployed and technology like p2p does help in the process. That being said, I’d still contend that in the not so distant future bandwidth will be nearly infinite (or consume a trivial amount of opex). Just over a decade ago it cost $250K/month for 45Mbps between the US and Japan. Today, the prices I have seen are a small fraction of that price. Tomorrow, it will be even less and will have CDN services too.

    So – what do we do with this bandwidth? Personally, I’m looking forward to installing my HD video wall at home and being a part of an immersing, life-like social network.

  10. I agree with previous poster, this was a very tricky tactic of posting on GigaOm, redirecting me here, and then serving up an article devoid of anything of real value.


Comments have been disabled for this post