The Need for Speed: Over-Delivering Bandwidth
Questioning demand for higher bandwidth in an effort to understand the repercussions.

By: Bruce Bahlmann - Contributing Author (your feedback is important to us!)

Created: July 21, 2005

This paper is the product of Broadband Market Research which is available from Birds-Eye Network Services.

In an effort to understand the “need” for speed, this article looks at the basis of services provided over broadband and how frequently they are used. It is believed that knowing such information will provide clues in proposing some kind of intelligent (based on market research) estimate of what consumers “actually need” versus what service providers and their vendors think or are made to believe consumers need. Such information will also be used to pose the counter view point offered up in this article surrounding the idea of ”over-delivering bandwidth” which will be discussed in terms of the repercussions such an offering raises as well as what the price might be if service providers guess wrong. The inspiration of this article came out of reading numerous articles written about the interest and potential profits that lie on the horizon of high bandwidth services (100Mbps and beyond) and perhaps just thinking too much about the subject in general. I would like to thank Peter Sheedy of Sedna Services for his contributions, patience, and similar passion to my own regarding the subject matter of this article. 

Hype, Reality, or Insanity? 

High bandwidth services (100Mbs and beyond) are not rocket science. In fact, JCOM and Cablevision are offering 100Mbs services to residential customers today using proprietary technology such as that from Narad Networks which works above 1GHz. Verizon is also offering a FiOS fiber to the home service that could easily ratchet up to 100Mbps and beyond. Only proprietary 100Mbps services have enormous challenges ahead before they can become ubiquitous (see Birds-Eye.Net article about Speed Bumps). Certainly Consumer Electronics (CE) is going to play a much more critical role over the proprietary solutions we are seeing deployed today. What is more likely to happen is that CE will finally settle into the driver's seat and come up with something revolutionary to today’s service provider specific access devices. Perhaps such a revolutionary device could be a universal modem that will enable consumers to connect to any service provider network and expand their needs for high bandwidth independently of which operator they choose to deliver such service. Or force service providers to enable some kind of pod-expansion module that will enable this universal modem to work with one or more service provider networks (DSL, DOCSIS, FiOS, etc.) 

Before you start rubbing your hands together and dreaming about all the possibilities, a number of questions surface when you think about high bandwidth services beyond the notion of how cool it would be to have a 100Mbps pipe into your home. For example, how much bandwidth do consumers really need? Why would operators offer such a service? Is there any reason that broadband providers should think twice about deploying such a service? To answer such questions you to really think about the “need” for speed based on consumer need for entertainment and access. 

How Much is Enough or Is There No Such Thing? 

Having a connection to the Internet isn’t enough these days, or at least so it seems. Like Moores Law predicting a doubling of processor speeds every 18 months, bandwidth availability to the home is also increasing at an equally alarming rate. What is driving this increase? Are consumers requesting more bandwidth or are content providers demanding more? Is there really sufficient content out there to justify the increase in bandwidth? 

In a previous article we examined the evolution of residential bandwidth offerings. Since 1996, bandwidth offered to the home has grown 1,000% and the price per 1kbps of this bandwidth has dropped 90%. Taking these previous trends into account it is predicted that by 2011 the retail price per 1kbps will be around 0.0273 cents which amounts to about $27.30/month for 100Mbps. While that doesn’t sound too bad, one has to ask the question, how much do consumers really need? What if unlike computers and processors, bandwidth has a ceiling? If the demand for bandwidth were to level off the resulting bandwidth glut would send residential data service prices into a free-fall. In an effort to get our arms around bandwidth demand, Table 1.0 describes bandwidth consumption rates for popular media types regularly consumed today as well as an up-and-coming High Definition Television (HDTV) service.  

Table 1.0 Streaming Media Bandwidth Consumption Rates (as of 5/2005) 

Note that none of these multimedia types are no more than 8Mbps. Meaning that if you want to receive a HDTV video to a home all you need is about a 8Mbps pipe or larger. Today’s broadband services are commonly offering between 4-6Mbps and in some cases offering 16Mbps or even 100Mbps which means that the gates are open for any service who can arrange to stream video down to consumer homes can offer video services – including switched digital video offerings. 

In Table 1.0, other forms of media on the Internet were not listed (such as web pages, emails, instant messaging, etc.) While clearly such forms of media consume bandwidth, they do not stream (or represent long durations of bandwidth consumption). With this information in hand, the next question asked was how frequently do consumers request these types of services or content? Table 2.0 provides some details around annual consumption (take) rates for various multimedia that have been drawn from related numbers who sources are hyperlinked. 

Table 2.0 Media Consumption Frequency (as of 1/6/2005) 

Numbers based on*:

$14.95 – average retail cost of DVD (estimated)
70 million homes in US with a DVD player
$3.95 – average retail cost of DVD rental (estimated)
240 million homes in US with one or more computers 

Determining consumption rates provides a framework for calculating the scope of what consumers need from a telecommunications standpoint. Meaning, if you have all these various forms of multimedia and one knows the general tendencies of consumers to experience or obtain them, you can determine (at least for the purposes of this article) a theoretical minimum probability for consumers to experience all these simultaneously. Such a framework can create a basis for how to package bandwidth based on what services a package might support (or not support by virtue of what the bandwidth will not allow). The alternative to this reasoning is to package bandwidth based on what a consumer actually wants, needs, or is willing to pay for or alternatively you can simply base your packaging on what your competition is offering. The repercussions of not observing consumer need should not be overlooked – this will be discussed later.

Table 3.0 Media File Sizes (Penn State University ResCom Dept.) 

Combining Table 2.0 and Table 3.0 one can arrive at a reasonable projection of a broadband customer’s annual apatite for bandwidth (see Table 4.0). Such information would be helpful in creating service tiers or if you are selling bandwidth by the bit, you could set your watermark sufficiently high enough to capture average consumer tendencies and then charge high users a premium. In this way, you keep abreast of take rates and average consumptions in order to provide guidance to your size and pricing tiers. 

Obviously the narrower the pipe (lower the bandwidth), the fewer the services your broadband customers can legitimately receive (at least optimally). With that said, the larger the pipe (higher the bandwidth), the more services each customer can receive. There is a saying in broadband that if you build it they will come. Only if you build out large bandwidth pipes to your broadband customers without the readily available content and services to fill that bandwidth, someone else (or some other company) will be happy to fill that spare bandwidth for you – thus limiting your captive, patient audience for new services as you roll them out. See article, Lead with Content not Bandwidth for detailed explanation of the importance of readying content offerings prior to boosting bandwidth speeds to residential customers. 

Table 4.0 Media Consumption Frequency (as of 1/6/2005) 

Based on the average consumption of media outlined in Table 4.0, a typical home will consume about 2,664GB per year, or around 222GB per month. If we say that one of the 3.1 televisions found within the home is High Definition (HD) the bandwidth consumption for roughly 520 hours of television viewing increases by (520 hours * (8Mbps – 3.5Mbps)) or 1,053GB. As a result, our new average consumption of media of a typical home (with one HD) is about 3,717GB per year or around 310GB per month. Such numbers assume transport of all these services over a standard Internet Protocol connection so all services would essentially be IP services. 

Absence is the Mother of Invention 

Show me a legitimate need that consumers have out on the Internet and I’ll show you a dozen or so companies in the works or activity serving that need. Meaning, if there is something people need out on the Internet, someone or some company is actively looking to sell them what they need. Internet time was coined by Jim Clark in a book about him called the New-New-Thing, and it is the Holy Grail for virtual brick and mortar businesses. Internet time evolved as the result of Jim Clark seeing first hand how quickly things materialize on the Internet. The Internet holds steady on the bleeding edge where trial and error can quickly change from rags to riches or possible fortunes lost. While this is not always the healthiest place for successful companies to live, it is where innovation and creativity come together to offer consumers what they want when they want it. Think of it as a place where instead of saying “You Will” the Internet says “You Can Today”.  

For slow moving companies, the Internet is this pesky little enemy that is constantly vying for your customers. Only the Internet doesn’t lambaste you or your customers with advertisements or go public big time so as to draw attention to it self like advertise during the Super Bowl, on local radio, or work with local retailers. Instead it works quietly in the background noise where it is more than happy just chipping away and not really bothering anybody. I equate the operations of your average Internet company to the following: 

If you go to a large city you will hear this constant hum. The hum is the resulting sum of all the sounds of the city. In this environment, think of an Internet company as a single jet flying overhead. In the city, not a lot of people walk around looking up – rather they look where they are going so as to avoid potholes, curbs, and other people. Unless you specifically stop, look up, and listen for it – an Internet company jet just zips past without notice. Over time it may reach a point where you notice it – however at that point it has risen above the hum and becomes unlike the obstacles you are used to avoiding. This is the kind of obstacle that you might kick yourself for not listening for or more carefully looking for in the past. 

Packaging IP Services 

Ideally, if you never wanted someone or some company to sell services over the top on you, you would never offer a broadband service greater than 6 Mbps. At such a bandwidth, customers are hard pressed to receive HDTV services and any streaming digital video service might still be border line at 6Mbps. Moving beyond 6Mbps to 10 or even 16Mbps has a number of consequences. Any bandwidth service 11Mbps and beyond can support the 3.1 simultaneous digital video streams, a VoIP stream, and reasonably good Internet service when all televisions are not being used. A service offering 16Mbps or higher can support 2.1 digital video streams, 1 HD video stream, a VoIP stream, and reasonably good Internet service when all televisions are not being used. 

Note that the above cautionary service speeds are best case. In reality, most services are not 100% efficient and to be able to comfortably view a 3.5Mbps video stream you probably need plenty of overhead bandwidth available to safely experience such services without artifacts or other quality hiccups. If such services are not delivered using Quality of Service (QoS) technology, the overhead bandwidth required to safely enjoy such services may be in the order of 30-50% just to approach something viewable. Meaning to obtain a 3.5Mbps MPEG stream, one may need a broadband service offering of 4.55 – 5.25Mbps worth of available best effort bandwidth. The more safety provided, the higher the likelihood that a service could be streamed down a best effort network un-impacted by contention or other network obstacles. At the speed of 2x (32Mbps) or 3x (48Mbps) a best effort service has enough headroom to more easily tackle contention obstacles in its way – perhaps even to the point where the value add of QoS becomes diminished by the abundance of available bandwidth (see QoS article). The higher the bandwidth offering of services, the less critical the need for QoS – and thus the more attractive alternate (over the top) provider services will appear to consumers from a quality and price perspective. Raising bandwidth to consumer homes without readily available content to sell to consumers and without fear of over the top exploitation of this bandwidth cannot represent a calculated business move to grab the money laying on the table for such services. In the end, consumers are not buying bandwidth. Consumers are buying services and those broadband service providers that judiciously raise their service bandwidth only in step with readily available content they can deliver to their customers are assured to survive the eventual commoditization of bandwidth. On the other hand, those who reach for the money on the table by offering high bandwidth services without readily available content open themselves up for an onslaught of Internet services hungry for the opportunity to finally rise above the hum.

Check out these other Birds-Eye.Net papers/products regarding this article:

• Bandwidth Contention and Profits - Combating forces working against efficient, profitable delivery of quality data services.
• Broadband Market Research - Great place to start learning about broadband
• Impact of Content Demand on Bandwidth Management - Addressing the source of broadband's network capacity depletion
• Lead with Content and Services, not Bandwidth - What is more valuable to consumers – a pipe or the water that flows through it?
• Limiting Top Talkers on a Budget - Employing restrictions of subscriber bandwidth usage without packet shaping.
• Speed Bumps on the Battlefield of Broadband (price) Wars - Obstacles to speed increases and the ongoing expansion of broadband.
• The Future of QoS - It will not be the differentiator that service providers are looking for
• The "Need" for Speed: Over-Delivering Bandwidth - Questioning demand for higher bandwidth in an effort to understand the repercussions.

Can Birds-Eye.Net help you or your Company?
Receive your Birds-Eye.Net articles and white papers hot off the presses by adding our RSS feed to your reader.

(C) Copyright Birds-Eye.Net, All rights reserved.
It is against the law to reproduce this content or any portion of it in any form without the explicit written permission of Birds-Eye Network Services, LLC. Federal copyright law (17 USC 504) makes it illegal, punishable with fines up to $100,000 per violation plus attorney's fees.