A lot, but not as much as one might think.
If you buy rather than lease hardware, the hardware's not too bad. Operatings systems and software that'll stand up to the heavy loads of these kinds of boards is pretty expensive and if you do it in-house (the only way to do it on a system with heavy loads), there's an ongoing expense for a programmer. There's message board software out there available for free or cheap, and some of it's really good, but none of it that I've seen in that category is very hardware/bandwidth-efficient and quickly becomes non-navigable with a lot of traffic.
I've got a pretty good idea of what I think SI is worth but won't say because *if* it ever goes up on the block, I intend to be a bidder. What I think it's worth is likely far less than what INSP thinks it's worth but also likely more than anyone else does if they're looking at it from only an ROI perspective. It's worth more to me because of the value I put on the community and history.
Although it might be worth more to anyone who doesn't understand these things under the hood well enough to know that this one could really break the bank in hardware, bandwidth and code-monkey costs if it ever returns to its old traffic levels.
The software really makes the difference in bandwidth and hardware expense. My software could run this site very well at current traffic levels on a T1 with 3 (relatively potent) machines yawning most of the time. The way SI is written, though, it would take about 4 T1's and a decent sized server farm full of heavy-hitter machines.
If you use some of the free or cheap software out there to run a site, you could do it on one machine with a 512kbps connection but would run out of bandwidth and hardware at just a few hundred posts per day.
And to prove that I don't know it all, look at iVillage. That's probably the worst approach I've ever seen to message board software. If you would've asked me at the outset, I would've sworn on a stack of ASP/SQL How-To books that they could never run at sufficient speed or attract a sufficient number of users with that approach to message board software. But they're big. REAL big. I'm sure they're using far more hardware and bandwidth than a comparable ASP/SQL setup would use, but my guess is that even ridiculous hardware/bandwidth costs to them are not a bank-breaker. An even bigger challenge is getting people to put up with that interface, but they've apparently struck a chord so strongly that the interface could be an Etch-A-Sketch and they'd still get the traffic.
I shudder to think what they could be/do with industrial-strength software really designed for the user. |
