Friday, November 15, 2002
The Flat Rate Breaking Point
Posted by Jason Dunn in "THOUGHT" @ 04:40 PM
Nicholas Negroponte's book, Being Digital (affiliate link) was perhaps the most powerful influence on my view of the digital world in the '90s. He's published an article on "Being Wireless" that is quite interesting - and he speaks of an evolution towards a type of wireless ecosystem where access points are spread out like lilly pads on a pond. It's a great concept, but if it's a little too optomistic? I have a few thoughts on this concept that I wanted to share.
Here are a few quotes (though I suggest you go read his entire article):
"But 802.11 systems — now available in a variety of flavors, including 802.11b, widely known as Wi-Fi — do not stop at the walls of your home. Depending on the intervening materials, a vanilla Wi-Fi can radiate more than 1,000 feet. Since I live in a high-density area, my system reaches perhaps 100 neighbors. I do not know how many use it (totally free) — frankly, I do not care. I pay a fixed fee and am happy to share.
Because further down the street, beyond the reach of my system, another neighbor has put in Wi-Fi. And another, and another. Think of a pond with one water lily, then two, then four, then many overlapping, with their stems reaching into the Internet. (Credit for the water lily analogy goes to Alessandro Ovi, technology adviser to European Commission president Romano Prodi.) Look at the numbers: 3G, in its most generous projections, will deliver data speeds of 1 megabit per second — in two years. Today, Wi-Fi commonly provides 11 megabits, offering up to 54 megabits. Which standard do you think will be adopted?"
Overall, this is a brilliant concept for people living in crowded metropolitan areas - it could seriously redefine the connected nature of how people live if they're constantly in a Wi-Fi saturated area.
Does Free Have Limits?
However, I'm not so sure about the "free for everyone" - there's an economic factor here that Negroponte isn't mentioning: flat-rate has a breaking point. Let me give you an example. My cable modem costs me about $35 CND a month, and I'm online many, many hours a day. I use up far more bandwidth than my next door neighbour (or likely anyone on my block), and he pays the same amount as me. That's the nature of flat rate: the ISP might lose money on me, but he'll make money on my neighbour because while I use up more bandwidth than average, my neighbour will not. When I go eat lunch my bandwidth use drops to zero.
Now let's say that this "lilly pad" concept takes off - what happens? We have previously untapped capacities being maxed out across the board. I may be off eating lunch, but 10 other people are using up the bandwidth that would normally be going elsewhere on the network. Cable modems are notorious for being slow if they're in an area where a lot of heavy users are clustered together - bandwidth is not yet unlimited. There are limits going in, and limits going out. So what happens when the ISPs need to install fatter pipes to keep up a certain quality of service? Costs to end-users go up. My $35 CND a month might double to $70 if my whole neighbourhood was using it.
Do We All Want to Share?
And let's not forget basic human greed: I like my bandwidth fast! If I'm downloading a large file and I'm getting 10 KB/s because twenty other people are using up my bandwidth, the bandwidth I'm paying for by the way, guess what will happen? Bye-bye free access for my neighbours. Sharing excess capacity is one thing, but if sharing my access interferes with my ability to get work done, guess what goes?
But truth be told most people are not like me - my parents have a cable modem but I'd be surprised if they used up 0.01% of the bandwidth they potentially could in a 24-hour period. There's a lot of bandwidth out there not being used, but what's the breaking point? Or is there so much excess capacity that this is a moot point?
Brother, Can You Spare Some Bandwidth?
A possible solution to this would be some sort of flexible sharing structure where a client (laptop, Pocket PC, etc.) runs software to sniff out wireless access and request permissions from the owner to use it. Perhaps the router software would be smart enough to understand who the primary user is (the owner) versus the "borrower" and allocate bandwidth effectively. If I have 500 KB/s downstream, it's rare to have a server dish up more than 300 KB/s anyway - so cap me at 300 KB/s. That leaves 200 KB/s to be doled out among the wanderers with Wi-Fi cards. If I'm not using the bandwidth (I must be sleeping at this point), it opens the flood gates and shares it with others. When I come back online, the shared bandwidth narrows again.
We all like to get paid, and there's a possibility here to make everyone happy - what if, as part of this flexible sharing structure, there was a billing component built in? The ISP would turn each client into a value-added reseller of bandwidth. When Joe Somebody comes needs a connection, he fires up his laptop and goes looking for access points. He finds one, punches in his credit card number (or pays via PayPal), and rents the use of that access point for a couple of bucks. 50% of that rental fee goes towards the ISP who provides the pipe, 50% goes towards the customers who's connection is being used. Everyone wins. The control over this rental process would be left up to the user - if I didn't want to loan out my connection, I wouldn't have to.
Security would obviously be an issue, but it's not so complex that it couldn't be solved in short order. We could all become bandwidth barons in fairly short order.
Now where did I put that regal staff?
Here are a few quotes (though I suggest you go read his entire article):
"But 802.11 systems — now available in a variety of flavors, including 802.11b, widely known as Wi-Fi — do not stop at the walls of your home. Depending on the intervening materials, a vanilla Wi-Fi can radiate more than 1,000 feet. Since I live in a high-density area, my system reaches perhaps 100 neighbors. I do not know how many use it (totally free) — frankly, I do not care. I pay a fixed fee and am happy to share.
Because further down the street, beyond the reach of my system, another neighbor has put in Wi-Fi. And another, and another. Think of a pond with one water lily, then two, then four, then many overlapping, with their stems reaching into the Internet. (Credit for the water lily analogy goes to Alessandro Ovi, technology adviser to European Commission president Romano Prodi.) Look at the numbers: 3G, in its most generous projections, will deliver data speeds of 1 megabit per second — in two years. Today, Wi-Fi commonly provides 11 megabits, offering up to 54 megabits. Which standard do you think will be adopted?"
Overall, this is a brilliant concept for people living in crowded metropolitan areas - it could seriously redefine the connected nature of how people live if they're constantly in a Wi-Fi saturated area.
Does Free Have Limits?
However, I'm not so sure about the "free for everyone" - there's an economic factor here that Negroponte isn't mentioning: flat-rate has a breaking point. Let me give you an example. My cable modem costs me about $35 CND a month, and I'm online many, many hours a day. I use up far more bandwidth than my next door neighbour (or likely anyone on my block), and he pays the same amount as me. That's the nature of flat rate: the ISP might lose money on me, but he'll make money on my neighbour because while I use up more bandwidth than average, my neighbour will not. When I go eat lunch my bandwidth use drops to zero.
Now let's say that this "lilly pad" concept takes off - what happens? We have previously untapped capacities being maxed out across the board. I may be off eating lunch, but 10 other people are using up the bandwidth that would normally be going elsewhere on the network. Cable modems are notorious for being slow if they're in an area where a lot of heavy users are clustered together - bandwidth is not yet unlimited. There are limits going in, and limits going out. So what happens when the ISPs need to install fatter pipes to keep up a certain quality of service? Costs to end-users go up. My $35 CND a month might double to $70 if my whole neighbourhood was using it.
Do We All Want to Share?
And let's not forget basic human greed: I like my bandwidth fast! If I'm downloading a large file and I'm getting 10 KB/s because twenty other people are using up my bandwidth, the bandwidth I'm paying for by the way, guess what will happen? Bye-bye free access for my neighbours. Sharing excess capacity is one thing, but if sharing my access interferes with my ability to get work done, guess what goes?
But truth be told most people are not like me - my parents have a cable modem but I'd be surprised if they used up 0.01% of the bandwidth they potentially could in a 24-hour period. There's a lot of bandwidth out there not being used, but what's the breaking point? Or is there so much excess capacity that this is a moot point?
Brother, Can You Spare Some Bandwidth?
A possible solution to this would be some sort of flexible sharing structure where a client (laptop, Pocket PC, etc.) runs software to sniff out wireless access and request permissions from the owner to use it. Perhaps the router software would be smart enough to understand who the primary user is (the owner) versus the "borrower" and allocate bandwidth effectively. If I have 500 KB/s downstream, it's rare to have a server dish up more than 300 KB/s anyway - so cap me at 300 KB/s. That leaves 200 KB/s to be doled out among the wanderers with Wi-Fi cards. If I'm not using the bandwidth (I must be sleeping at this point), it opens the flood gates and shares it with others. When I come back online, the shared bandwidth narrows again.
We all like to get paid, and there's a possibility here to make everyone happy - what if, as part of this flexible sharing structure, there was a billing component built in? The ISP would turn each client into a value-added reseller of bandwidth. When Joe Somebody comes needs a connection, he fires up his laptop and goes looking for access points. He finds one, punches in his credit card number (or pays via PayPal), and rents the use of that access point for a couple of bucks. 50% of that rental fee goes towards the ISP who provides the pipe, 50% goes towards the customers who's connection is being used. Everyone wins. The control over this rental process would be left up to the user - if I didn't want to loan out my connection, I wouldn't have to.
Security would obviously be an issue, but it's not so complex that it couldn't be solved in short order. We could all become bandwidth barons in fairly short order.
Now where did I put that regal staff?
