Why do some ISP's have bandwidth quotas?

Joe Greco wrote:
>> Technically the user can use the connection to it's maximum
>> theoretical speed as much as they like, however, if an ISP has a
>> quota set at 12G/month, it just means that the cost is passed along
>> to them when they exceed it.
> And that seems like a bit of the handwaving. Where is it costing the
> ISP more when the user exceeds 12G/month?
> Think very carefully about that before you answer. If it was arranged
> that every customer of the ISP in question were to go to 100%
> utilization downloading 12G on the first of the month at 12:01AM, it
> seems clear to
> me that you could really screw up 95th.

First, the total transfer vs. 95%ile issue. I would imagine that's just a
matter of keeping it simple. John Q. Broadbanduser can understand the
concept of total transfer. But try explaining 95%ile to him. Or for that
matter, try explaining it to the average billing wonk at your average
residential ISP. As far as the 12GB cap goes, I guess it would depend on
the particular economics of the ISP in question. 12GB for a small ISP in a
bandwidth-starved country isn't as insignificant as you make it sound. But
lets look at your more realistic second whatif:

Wasn't actually my whatif.

> 90GB/mo is still a relatively small amount of bandwidth. That works
> out to around a quarter of a megabit on average. This is nowhere
> near the "100%" situation you're discussing. And it's also a lot
> higher than the 12GB/mo quota under discussion.

As you say, 90GB is roughly .25Mbps on average. Of course, like you pointed
out, the users actual bandwidth patterns are most likely not a straight
line. 95%ile on that 90GB could be considerably higher. But let's take a
conservative estimate and say that user uses .5Mbps 95%ile. And lets say
this is a relatively large ISP paying $12/Mb. That user then costs that ISP
$6/month in bandwidth. (I know, that's somewhat faulty logic, but how else
is the ISP going to establish a cost basis?)

That *is* faulty logic, of course. It doesn't make much sense in the
typical ISP scenario of multiple bursty customers. It's tricky to
compute what the actual cost is, however.

One of the major factors that's really at the heart of this is that a
lot of customers currently DO NOT use much bandwidth, a model which fits
well to "12G/mo" quota plans. It's easy to forget that this means that a
lot of users may in fact only use "500MB/mo". As a result, the actual
cost of bandwidth to the ISP for the entire userbase doesn't end up being

If that user is only paying
say $19.99/month for their connection, that leaves only $13.99 a month to
pay for all the infrastructure to support that user, along with personnel,
etc all while still trying to turn a profit. In those terms, it seems like
a pretty reasonable level of service for the price. If that same user were
to go direct to a carrier, they couldn't get .5Mbps for anywhere near that
cost, even ignoring the cost of the last-mile local loop. And for that same
price they're also probably getting email services with spam and virus
filtering, 24-hr. phone support, probably a bit of web hosting space, and
possibly even a "backup" dial-up connection.

That makes it sound really nice and all, but the point I was trying to
make here was that these sorts of limits stifle other sorts of innovation.
My point was that cranking up the bandwidth management only *appears* to
solve a problem that will eventually become more severe - there are going
to be ever-more-bandwidth-intensive applications.

That brings us back to that question of how much bandwidth should we be
able to deliver to users, so the $6/user is certainly relevant in that

... JG