RSS.

Not to be supportive of adobe per se (they are miserable too) but that is why dng was developed. Raw is very hard to support from an archival perspective.

1) I guess it goes down until it can be fixed under warranty (same or next day depending on purchase option). Redundancy is expensive. What happens when your single instance of AWS goes down with an "oops amazon is having problems with a datacenter" message?
2)Good job, you have identified why Netflix uses AWS.
3) Reserved instance is cheaper, but at that price still more than a dedicated server and the server typically comes with a 3 year warranty and will likely last past that (Dell will warranty for 6 years). Assuming it only lasts 3 your cost for running on AWS is nearly 3 times higher even when figuring in an improved warranty and OS licensing. I concede that short duration projects or very spiky loads are a great use for the cloud, but long running relatively even loads simply don't make sense form a cost perspective, nevermind the fact that you now lose access to your database if your wan connection goes down (unless you build out multi-wan, but there is yet another expense).

Use OSPF and use pricing as one of the variables for cost calculation? Wouldn't take a rocket scientist and pretty sure any decent sized network does this already... We are't talking spot market here - most of these costs are negotiated in long-term contracts, but no reason we couldn't design it like the energy markets (though not sure you would want to).

Try it yourself. You might find out that to be successful at it requires skill and education. I have family there that are one of two families left in a several hundred mile radius that are still farming successfully. All the rest gave up or gave out. Between droughts, harsh winters and fluctuations in feed prices it isn't as easy as watch some cows munch grass, and yes, they are educated and have dedicated fiber running straight to the farm that far surpasses the quality of dsl I can get in the city here in Florida.

Right, and how is the firmware on the drive for your non-magnetic media holding up after that EMP blast? You did remember to load a copy of the firmware onto a disk too, right? Oh, and the bios for the computer you were planning on restoring to, and the hard-drive firmware and other various chipset firmwares? I think come an EMP blast you had better set the computer aside and know how to be a dirt farmer before you starve. Even if you get your own files restored it is unlikely you will be able to do much else unless you plan on helping the telco reprogram all their equipment to get the network back up etc. In the meantime, you starve.

Car analogy: I told the used car dealer to stop selling that garbage and just send all his vehicles to the dump. I mean they were all from like 2007 or before! I mean seriously, who uses a car that old (except for all the retro ones that were sold up until 2012 - and those suck too. They aren't hip at all)? They don't have the latest rear view cameras and other safety equipment or anything. It is no secret if you buy the after market warranty you can get your crappy old car fixed, but if you don't it isn't my problem you can't get parts when you need them because you are a dumb poopy pants. I throw everything away because there is a newer model that surely must be better because new and shiny!

because the spf records don't pass but the recipient recognizes the sender?

That is a terrible policy. I spent a long night at an office of a fortune 500 company for that very reason. They didn't see any reason to apply bios patches because they were just to add support for newer hardware, not to fix any sort of vulnerability. Fair enough. Several years went by and their terminal server had a processor go finicky on them. They determined the available spares included processors that were compatible. I asked "has the bios been updated to support the newer processors?" I was assured that they do regular patching and it would not be a problem. I arrive on site, install the new processors and get no post. A bit of troubleshooting and we determine it doesn't recognize the processors because the bios was out of date. Really long story shortened - we had to shutdown another server, pull the processors, install them in the problem server, boot, patch the bios, shut down move the processors back in the donor server, and then reinstall the new processors. Of course this was in a server room that was an overstuffed shoe box so a number of acrobatics were required to get the servers extended to a point they could be worked on.

So what should have been a 10-15 minute processor replacement ended up causing several hours of downtime and the unscheduled shutdown of another server.

Don't be lazy!

That said, as someone else stated, I usually wait a couple months to patch (especially HP) unless it is considered a critical issue or I have a straightforward fail-over plan. HP has screwed my arrays etc. more than once with their quality updates.

No, you don't understand. All the little internet connected things in your life like your thermostat with infrared sensor and tv with camera and xbox with 3d imager and phone with gps and toilet with butt activated hemorrhoid sensor all send their little bits of data back to the big intelligence in the cloud. This way the great data architects of Fort Meade know you need some anal cream, a diet, and some new pants. They might also recognize that you are a danger to yourself if you continue to play WoW. But if you stop playing you might be upset about your surroundings and be a danger to others, so you get a new online friend to help you play even more hours each day. I think the Internet of Things is quite Intelligently Designed. In fact, I think everyone else who supports it should all spread the word by using a hashtag for intelligently designed internet of things #IDIoT

It's already in use by the administrative share.

This is 60 minutes we're talking about here... don't get too demanding!