I’m attending WebVisions here in Portland Thursday and Friday. Good stuff so far, and very different than those mega-conferences I’m used to at Moscone Center in San Francisco.
Rather than let you pine for content, here’s a taste of something new. Some of you may know that I made a guest apperance on Web Worker Daily last month. I had planned to contribute sporadically, and I had put together a piece on bandwidth policing.
However WWD’s kind editor, Judi Sohn, has a glut of content, and apparently GigaOm, the parent company of WWD has recently reworked its contributor policy to require a contract.
All this means I have an extra post waiting to see the light of day. Here it is. Enjoy.
Bandwidth is the web worker’s lifeblood. It can never be too fast or too plentiful.
Although many web workers frequent office shares, coffee shops, and other public wi-fi locations, most of us have a home office, which means we pay an ISP for bandwidth.
Advances in broadband and its increasing reach, 77% of all US home Internet users and 55% of all Americans have a broadband connection, have put a strain on broadband service providers. So, as more people watch streaming video and download media, ISPs have begun to worry that their billion-dollar infrastructure investments may not be enough to handle the demand (read diminishing return on investment). This has led to doomsday predictions like that of AT&T’s Jim Cicconi.
Enter bandwidth policing.
Although they’ve been there all along, more recently announced caps on bandwidth and restructuring of the pricing models to pay-per-use have web workers in an understandable uproar. Comcast’s plan to slow down the connections of people it deems as “overutilizing” their service seems especially unfair to web workers.
Web working requires fast, dependable bandwidth, and there’s no way for an ISP to tell determine the intented use for a packet. Packets from iTunes could be songs for entertainment or podcasts for work. Streaming video from YouTube could be the latest Rick Roll or a potential client’s video demonstration. A large mutli-gigabyte download could be a movie or it could be a software installation.
So, web workers who need fast connections to, you know, work, could be penalized for their usage. You might argue that pay-per-use would be fair, but web working doesn’t come with an enterprise budget. A rise in core costs negatively impacts the web worker’s business because it must either absorbed into the business or passed on to customers. Neither is ideal.
The same problem applies to people like me who telecommute for a large company. My employer offers telecommuting to me as a benefit. However, say my ISP raises my rates because they see me as a serial bandwidth hog. I download software, watch recorded video from conferences, host and attend web conferences and generally use a lot of bandwidth to do my job. Plus, I do it over a VPN connection, so each packet I send and receive has an additional overhead.
What happens when my cost rises? Just like freelance web workers, I’m forced to decide between absorbing the additional cost or passing it on, in this case, to my employer. If I pass it on, I’m running the risk that my company will see the telecommuting benefit as too expensive. In an uncertain economic climate, benefits all come under scrutiny, especially since my employer has perfectly good office space I could use. So, I’m the loser in this game.
It seems unlikely that issues like bandwidth policing and net neutrality will resolve quickly, so for the moment, we have choices. Some ISPs, like COPOWI, promise net neutrality, but since they use the existing infrastructure of a major ISP, their traffic would be subject to bandwidth policing, leading ultimately to the same dillemma.
For now, if your ISP is threatening to police your traffic, and you think you might be affected, vote with your feet and find a new ISP. You might consider mobile broadband as an alternative; several web workers I know have gone exclusively mobile because of speed (comparable to residential DSL) and convenience. Most providers offer an unlimited bandwidth option and prices are comparable to wired broadband.
Mobile broadband use is on the rise, and providers are more likely to keep it affordable and unpoliced as they lure in new customers. Because the market is less mature, you can probably expect a few years before diminishing returns hit.
When they do, let’s hope there’s another recourse, or that bandwidth policing has been squashed in favor of a more equitable solution.
“A rise in core costs negatively impacts the web worker’s business because it must either absorbed into the business or passed on to customers. Neither is ideal.”
Welcome to the real world. Demand for petrol (or diesel) outstrips supply, so the price goes up to curb demand. Users have to absorb the cost or raise prices.
In fantasy land, you can get unlimited broadband cheaply, or even free. In the real world, some has to pay for cable, routers and other hardware, for men to go out and fix stuff when it breaks. it is perfectly reasonable to expect users to pay for the resources they are using.
“bandwidth policing has been squashed in favor of a more equitable solution.”
Whole schools of political thought have been built on 'equitable' distribution of limited resources. Unfortunately, most break down as people want to be more equal than others.
I live in the real world. Somehow I have a feeling that bandwidth hogs account for less than 10% of the total customers of any given ISP. Therefore, the other 90% under-uses and over-pays. ISPs don't want to go full pay-per-use because it would negatively affect their businesses.
I haven't seen any stats on over-usage and its affect on the connections of everyone else. My sense is there are very few times during a day that over-usage affects the average users.
I don't have a problem with pay-per-use, but it can't be selectively applied. If so, under-users should revolt.
Also, I'm not sure that demand for gas outstrips supply. Supply is controlled by a small number of producers, so it's difficult to measure true supply-demand due to constraints. The cost of a barrel of oil has risen $100 dollars in 8 years. Is that due to a 300-400% increase in demand?
I agree that equality breaks down when applied to limited resources.
“The cost of a barrel of oil has risen $100 dollars in 8 years. Is that due to a 300-400% increase in demand?”
Guess you never studied economics. Simple example of why that logic doesn't hold true. A company produces 10 doses of a life-saving drug for a disease. There are 10 very rich people with the disease. No excess demand, so the company charges the cost of producing the drug (plus some profit). If 11 very rich people suffer from the disease, the 10 doses suddenly become 'scarce'. The drug company can increase its charges ten fold (or even a million fold), until it becomes too expensive and one of those individuals drops out. Even a 10% (or 1%) increase in demand without a corresponding increase in supply means scarcity, and the price will increase until demand is forced to drop.
With petroleum, there's been increasing demand from China and India. There hasn't been a corresponding increase in supply. The price will increase, not in proportion to the increase in demand, but until it reaches a point where demand falls.
Quite simply, the price of gas will go up until it reaches the point where enough people can't afford it.
The same will happen with bandwidth. The price will rise until people stop using it or the extra income pays for more capacity to meet the demand.
WRT bandwidth, your model may be correct. The problem is that the ISPs want to recoup their sunk costs on infrastructure, and their models don't meet the increasing trends, i.e. bandwidth hogs make them nervous. So, they want to have their collective cake and eat it too by charging the under-users a flat rate and over-users a pay-per-use rate. Makes sense as a business plan, but I don't like that as a consumer.
So, your point is what? If you think I should pay-per-use as a seriel bandwidth consumer, then I assume you are an under-user who would also like to pay-per-use. If you are an over-user, why would you want to pay for use?
I don't really get what lesson you're trying to teach me 🙂
WRT to oil, I have an economics degree, and we're both grossly oversimplifying the market forces of oil to make a point. Suffice to say, we could cherry-pick pieces of the whole that prove our points all day long.
Here is Oz, most internet plans are limited/pay per use. I pay for a 1Gb limit per month with usage above that 'shaped' (slowed down to about dialup speeds) or I can buy extra at about $5 per Gb. Some plans offer extra 'off-peak' usage.
The economics can get complicated. I don't have a problem with businesses offering whatever terms they (and consumers) agree to. Slowing users, charging more or otherwise limiting peak use, or charging higher prices so they have a high enough capacity/user ratio and don't have to limit.
Tightening down on illegal (copyright violation) downloads may be generally beneficial, if anyone can work out how to do it.
OK, I get where you're coming from now. I guess if broadband had always been pay per use, then no biggie. But, since it's been “all you can eat” for so many years, I feel annoyed that they want to make me pay per use.
Your plan sounds way too restrictive for me, makes it tough to work from home.
“makes it tough to work from home.”
I've got two kids (6 and 3). THEY make it tough to work from home 🙂
Ah, that seems like it would hurt your productivity at home. I just found out Verizon is laying fiber to my neighborhood in Portland next year, so maybe by late 2009, I'll be happily paying a lot more for a lot more pipe. I pay for 8 mbs now, but rarely get more than 50% of that, even less over VPN natch.
Pay for as much bandwidth we use? But that's like paying the minutes I talk on my phone or the data I transfer when I use my phone's internet connection. I don't want to feel limited when I work at my computer. I understand why the internet companies might be worried but…can't they just get bigger bandwidth?
Sandra Millhouse | VPS unlimited bandwidth