This is the technical equivalent of: "Hey, I found $100 what should I do with it"? 255 addresses ... woo-hoo.

"What would you do with a Class C IP block?"

"I'll tell you what I'd do, man: two chicks at the same time, man."

You don't need a Class C to do nothing, man. Take a look at my cousin: he's on dynamic IP, don't do shit.

Sort of... I'm not sure what I'd do with 256 IP addresses without 256 public-facing servers...

I'd much rather have a few hundred servers and 1 IP address.

1 server with an extra $500 in RAM can do some pretty fancy stuff with virtualization.

It seems a bit closer to a quarter than $100.

Hrm, I've found it rather easy to obtain IP space from ARIN so this article seems a little humorous to me.

Reminds me of joining IBM Global Services a few years ago:

'The IPs that start with 9 are ours.'

'Sure? What's the full netblock you own?'

'9. All of it.'

If you lived in a small neighborhood (or some apartment buildings, if the walls aren't too thick) you could start a little ISP. Or you assign all of the devices in your house an IP address, and even break down the ranges (e.g. N to M are for kitchen appliances.)

Why would thick wall be a problem? (I'm not being sarcastic, I really don't know)

because he's probably talking about doing it over wifi, which happens to degrade fast if it needs to traverse thick walls.

Give it back. If you can't use it, you don't need it.

255 IPs is a drop in the ocean.

Companies with Class As (16,777,214 addresses) are the problem.

The legacy Class As really aren't a significant issue either. By the time you subtract out all the effort that the clawback itself would involve (nontrivial reconfiguration of quite a few major corporate and government networks), it would probably be a wash or quite close to it versus just trying to work on IPv6 -- which finally seems to be getting some traction.

At best I've heard a few extra months added at current burn rate estimates. Hardly worth the knock-down, drag-out fight that it would involve.

I suppose that would depend on what ISPs you're buying transit through.

Technically, the smallest block you can announce that won't be filtered is still /24, and many do, including networks that punch holes in their own aggregates, or their upstream ISPs' aggregates for redundancy, etc.

However, with the global BGP table coming in at just shy of 300,000 routes now, there's pressure to filter small prefixes more and more. There's kind of a de facto gentleman's agreement among the sorts of folks that are in NANOG that /22 should be the new de facto "smallest prefix" wherever possible. I really think getting your /24 announced and propagated (if it's truly a provider-independent block from a RIR, e.g. ARIN, and not just a Class C subnet from an ISP network) might be encumbered by hemming and hawing from routing administrators that are constantly keeping the dwindling capacity of existing routers to hold full BGP views - especially multiple full views from multiple peers - in mind. You can definitely do it, but the appeal of a /24 is slightly diminished in light of that.

Rent them out?

I would use one for my personal use and take the remaining 254 and offer reverse dns services.

devise a numeric code so that users navigating across the ip's are receiving an "important" message. the website navigation plays along with this "choose your own adventure" game.

Just so you can feel smug - a class C block gives you 254 usable IP addresses, not 256.

You could always build a 255 site SEO network and interlink them all for SERP purposes. Class C blocks are big for people building link farms in order to take over the top ten results for different search terms.

Google already checks if sites belong to the same c-block, so the link value is significantly diminished.

This would be trivial for Google to detect, no?