I'm sure caker will hate me for suggesting this... but, I regularly under-user my bandwidth. And I always have at least a little extra disk space available.

In the meantime, I notice that a lot of small websites get slashdotted and the like by overwhelming peak demand. What I think would be cool would be to set up a fairly sturdy proxy cache system, possibly linked in with larger existing systems (like this one: http://www.ircache.net/), and peered against each other.

Ideally, to set this up, I'd write all the tools to generate a simple squid config for those who wish to participate. And use some DNS round-robbin hacks to let people control how heavily they're donating resources (so you can, for instance, make sure you don't over-use your bandwidth, or try to use up more toward the end of your billing period).

Would anyone else be interested?

Anyone want to weigh in on what kinds of additional information they're interested in knowing?

I discovered this in my search for prior work in this direction:

http://www.archive.org/web/freecache.php

It's run by the Internet Archive, and seems to be a reasonable service.

The only problem I have to say, and this is a big one - make sure you watch your bandwidth usage. I started it up yesterday, and had burned through roughly 1/2 of my monthly bandwidth allocation by morning. Yet one more reason to support "how much I want to help" metrics in the system I'm talking about.

Neat model, though. Very easy for the end-user to take part of, but not explicit caching of all content.

Thought I'd add my 2p seeing as I haven't posted for a while...

Athough a nice idea it's probably worth considering that the prices for Linode are likely to be based on the assumption that many people use a lot less bandwidth than their maximum allocation each month, and assuming that Chris gets charged for the amount of traffic, although the full bandwidth will be available to all of us, the returns on running each Linode is likely to be less with significantly more people using more bandwidth.

I'm not saying that nobody should bandwidth share (it's none of my business what people do with their Linodes) nore am I saying that a few more people using their max bandwidth will have any bearing whatsoever on costs - but am just adding a point for discussion, feel free to tell me I'm talking rubbish

(Maybe I'm just stingy and love these prices! )

That concern is exactly why I started out with "I'm sure caker will hate me for suggesting this", and been wondering if he will weigh in on the issue.

I can only guess, based on the prices of extra bandwidth that he offers, that you're approximately right. But, given that he does allocate a specific amount of bandwidth per host, I'd hope that he has the numbers worked out so that he still makes money if people more-or-less use up the whole allocation. The difference in pricing plans largely accounts for the cost of the extra hardware, yet each includes more bandwidth. This isn't a heavily over-loaded VPS setup, and caker's done a good job of showing that his model is designed to really allow us each a decent little pool of "our own" resources on the cheap.

So, what is it, caker? Would you prefer we not conspire to use as close to 100% of our bandwidth as we can?

I don't think it would make much of an impact, for a number of reasons. The number of people who would do this would be small. HE bills me for MBits/sec (95th percentile), not MB transferred -- this probably wouldn't have much impact on throughput. On the other hand, ThePlanet bills me for MBytes transferred. It all averages out to nearly the same costs for me, so I say go for it.

-Chris