Mac Musings

You Will Pay More for Broadband

Daniel Knight - 2002.10.28 -

The Internet as we know it is under attack.

The Internet was first envisioned as a "Galactic Network" connecting computers around the world by J.C.R. Licklider of MIT in 1962. Computers would not only be connected to a global network, but also able to share data. Forty years later, we take that for granted.

The first connection between remote computers took place between a computer in Massachusetts and another in California using "a low speed dialup telephone line" in 1965. As noted in A Brief History of the Internet, it was obvious that the telephone network of the day was painfully inadequate.

This led to the planning and creation of ARPANET, a national military and research network designed to operate at up to 50 kbps - about the top speed of today's v.90 phone modems. The first computer was connected to ARPANET in September 1969, marking the birth of the Internet.

The second node was at Stanford, and by the end of the year the fledgling network had grown to four computers. In 1972, the Internet's "killer app" was created when Ray Tomlinson invented email, giving birth to user-to-user interaction.

ARPANET evolved into the Internet, which is a conglomeration of independent networks spanning the globe - some as small as two computers in a home and others linking tens of thousands of machines. To accommodate a great variety of hardware and networking protocols, TCP/IP was devised as the common language by which devices would communicate on the Internet.

When the personal computing revolution ignited in the mid-70s, the Internet already existed. When computers went corporate and began to be networked, the Internet was ready. Only today, 40 years after the "Galactic Network" was first envisioned, are we running up against the limitation it the current IPv4 protocol - and IPv6 has already been invented to take its place.

Home Users

For home users, the first connections were between one user and another, often with 110 baud modems. Later came dial-in access to university computers and bulletin board systems. And modems improved to 300, 1200, and then 2400 bps, made a big jump to 14.4 kbps, doubled that to 28.8, and managed some clever tricks to finally enable downloads at up to 56 kbps over phone lines.

At the same time, competition emerged. Cable television companies already had a high bandwidth infrastructure, and they slowly began offering cable Internet connections. The phone companies responded by developing DSL, which offered a much faster connection over existing phone lines than standard dialup modems.

Today DSL is a moderate success and often offers connection speeds in the 1.5 Mbps range - but cable trumps that with potential speeds in excess of 10 Mbps. And, unlike most dialup services, cable and DSL Internet access offer 24/7 connectivity for a flat monthly fee.

Bottlenecks

Phone modems were the first bottleneck for personal computer users. At every step in their development, the underlying bandwidth of the Internet was much higher, so modem users could generally be assured of uploading and downloading data at whatever speed their modems could sustain.

Cable and DSL changed all that. Although the average bandwidth of the Internet keeps improving, in most instances today the bottleneck is somewhere between the ISP and the server the user is connecting to. Maximum download speeds of 10 Mbps may sound impressive, but in reality such speeds are rarely achieved.

Some businesses such as Akamai have attempted to address this by distributing content servers, but anyone who has attempted to watch a Macworld Expo keynote address knows that today's Internet doesn't really have the bandwidth we need.

Bandwidth

And now the cable Internet industry, which controls an increasingly large portion of the installed base, is talking about caps and limitations and extra fees. You can read a lot more about this in The Death of the Internet, but here's a quick overview.

It looks like that is about to change. Some cable providers are already offering different connection speed/pricing packages. At present, these are mostly of the unlimited use variety. All the cable ISP does is put a ceiling on how fast you can connect.

They'd rather switch to a model where they can tack on additional fees, much as the phone company does if you go beyond your contracted number of local phone calls or your allotted number of monthly minutes. The cable companies see big bucks if they can bill you for using more than, say, 5 GB of bandwidth per month.

How much is 5 GB per month? That's 167 MB per day, 7 MB per hour, 116 KB per minute, and just under 2 KB per second. Bytes-to-bits conversion makes that 16 kbps - a pittance compared with the 1 Mbps and better bandwidth usually promised with cable connections.

That's a very low cap, and it's already being tried by Bell Sympatico. Sympatico calculates that a 20 kbps radio feed would need 70 hours to use up your monthly bandwidth allotment. That's 70 hours of low quality Internet radio with nothing left - no email, no chatting, no use of the Web, no movie trailers, no online gaming, no software updates.

Say you work at home and find 20 kbps audio adequate. You could listen for eight 8 hour days during the month plus 6 additional hours before you'd use your monthly bandwidth quota. And then they'd start sticking it to you at maybe $1 per gigabyte - maybe $5/GB. And your cable bill would skyrocket.

Bandwidth Hogs

The industry is already promoting the term "bandwidth hogs" for those who actually attempt to use more than the barest percentage of the service they are paying for. Many ISPs have already redefined "unlimited" use to mean limited. And they are using a pejorative label to get you on their side.

The claim you'll hear the cable providers make is that bandwidth hogs increase their costs and tie up bandwidth others need. They may be correct on the cost side, but the bandwidth is a straw man argument. According to Mike LaJoie, vice president for advanced technology at AOL-Time Warner,

"The way that the HFC (hybrid fiber coaxial) architecture works, we never run out of bandwidth. We can always split or do other things that will give us the bandwidth that we want, so it really ends up being a desire to provide the best and highest experience for our customers." (HD on VOD Searches for Resolution, Multichannel News, 2002.09.30)

If bandwidth is essentially unlimited, as LaJoie claims, then it is ludicrous to complain about bandwidth hogs. It's the equivalent of calling someone a throughway hog for spending too much time on the Interstate. The infrastructure is there to be used; using it as intended is not being a hog.

That said, if there are additional bandwidths costs incurred by ISPs, it is only fair to set specific bandwidth limits at various fees, much as the cell phone industry provides so many minutes per month at this rate, and so many more at the next higher rate.

It does the cable industry no good to call heavy Internet users bandwidth hogs, because it alienates those who could well be their best customers. Instead, the industry should differentiate between light, moderate, and heavy users and offer plans accordingly.

A Proposal

Let's say that XYZ Cable currently offers unlimited cable modem use for $40 per month. That rate applies no matter how much or how little a customer uses the service, much as a buffet charges one price no matter how much or how little food one eats.

XYZ has several options for creating service plans. They can offer all users the same access speed and limit the amount of data per month, limit access speed while offering unlimited data within that bandwidth, or create a hybrid plan that limits both connection speed and total bandwidth used.

Corporations seeking profits will gravitate toward the plan that costs them the least and allows them to charge the most. XYZ Cable would much rather give everyone the same access speed and not have to worry about that aspect of the business while socking users with a surcharge for every GB beyond a predetermined amount.

Customers seeking the most for their money will gravitate toward the plan that costs them the least and gives them the fewest restrictions. We want to overbuy on our local phone calls or cell phone minutes so we never have to worry about the high fees for extra calls or minutes. The same goes for Internet access.

Unfortunately, cable ISPs have a virtual monopoly on broadband at present, so XYZ can pretty much call the shots. They'd really have to make it bad before customers would consider switching to DSL or before someone decided to create a wireless broadband alternative. They have the upper hand, and they know it.

We need to convince the FCC, which has abdicated its role as the protector of the public interest here, and the cable companies that bandwidth caps and pricing plans must be realistic.

The Moderate User Plan

The plan for the moderate user should be enough to cover listening to Internet radio at a decent streaming rate (say 56-64 kbps) 40 hours a week with sufficient bandwidth left for software updates, Web browsing, email, and watching a few movie trailers.

Based on the Sympatico figure of 5 GB equaling 70 hours at 20 kbps, we're looking at something more on the order of 40 GB per month, not 5.

I would suggest as a general guideline that a moderate user plan provide no less than 100 GB of bandwidth per month. And it shouldn't cost any more than today's basic broadband plan, since it's far less than what most of us are allowed today.

The Heavy User Plan

For heavy users, the calculation is much simpler: 1 Mbps around the clock - somewhere around 350 GB per month. That's a fraction of the potential 10-30 Mbps bandwidth of a cable connection, and AOL-Time Warner has already told us that bandwidth is essentially unlimited.

I can't even estimate how much the cable company should charge for such a plan. This amount of bandwidth is allowed under most basic cable ISP plans today, and I have no idea what the incremental cost per GB is to the ISP. I am certain that it is well under a dollar, though, and probably an order of magnitude less.

Let's call it 10¢ per GB just so we have a figure to work with. If the moderate user plan provides 20 GB for $40, the additional 300 GB would probably bring the plan to $70-80 per month - and the heavy user would probably be more than happy to pay that if the alternative was a fixed bandwidth cap or a surcharge on every GB used.

In fact, very heavy bandwidth users might be more than happy to pay for a super user plan that provided 10x the bandwidth - if only the cable company offered it at a reasonable price.

The Light User Plan

For light users, maybe a 10 GB cap makes sense, especially if it could trim their monthly broadband bill by $10.

Implementation

They say you can never have enough bandwidth, so the cable industry stands in an enviable position. They have the advantage and need to run with it to keep DSL and wireless from cutting into their market, much as Microsoft needs to keep Linux and the Mac OS on the sidelines.

The following are suggestions more than proposals. I don't know all the ins and outs of the cable ISP business. These are the thoughts of a user and consumer advocate.

I am a happy cable modem user. I love the connection speed, the fact that it doesn't tie up my phone line, and the fact that I can be on the Internet instantly at any time of day or night.

I don't want to worry about things like connection speed or bandwidth used. I might be willing to pay more for even better service, but I wouldn't want to pay more to simply keep using the service at the level I already have. That said, I'm sure I'm the kind of user the cable companies want to charge higher fees. And I'd be willing to consider it if they let me run my own server and the price was fair.

Take care of us, cable companies, and we'll continue to work with you. Stick it to us, and we'll pressure the FCC to drop their hands off approach and get involved in regulating your virtual monopoly.

Further Reading