And the difference between a witch and a warlock?

Witches don't have wands.

Thank you for succinctly explaining episodes I, II, and III.

I believe the 8kw was the stat for a rack (42U with 4 of these units). The stat for a single 10U box was 2kw.

The point isn't so focused on how many processors you can fit in that space, it's power consumption.

The link you provide details 2 1100W PSUs. That's in 2U. The summary quotes just under 2000W for one 10U server. Just looking at that, you're running at 1/5th the power consumption.

Thank you, AC. This is exactly the kind of project/ideas was looking for.

Funny, but you really needed the link: http://xkcd.com/566/

ok, tried this yesterday and got 5 locations in a row. Being more of a history/info buff than a geography buff, I modified to: http://en.wikipedia.org/wiki/Portal:Featured_content instead.

He's claiming 82% dedup savings with this. That's roughly five times greater than what you credit.

Even with the price overhead, I'd still consider a solution like this because I can replicate all my data on one storage appliance more easily than implementing replication across X commodity servers. Yes, I like to spend money to make my life easier. :)

While I appreciate the sentiment, it only applies to Big, Irreplaceable (tm) things. Tape libraries would be an example. Office software would not. Even if you have software that's mandatory, there are other ways to mitigate risks. Clustered servers for fail-over. Replication. Alternate installs in the form of development and test environments. If Wine breaks in the middle of my big project, I may research the issue and debug the problem, but if I'm under a time crunch, I'm just going to move to a working machine. Yes, paying for support is one valid risk mitigation strategy, but it's far from the only one.

Who fears this? After all, to counter:

/etc/init.d/penguind stop

Lich is just the retirement plan for a long and successful Necromancer career.

My initial, gut response to this was sheer horror. They list exploit and target side-by-side! The only mention of a fix is that it's to be 'released soon', informing any malicious agents out there that now is the time to strike.

Reading the Wired article, the right thing was done. Big company was sitting on their hands, and now that publicity has been made, they're starting to move.

Wired did the right thing. But this summary, it's fear-mongering and bad journalism.

An interesting analysis. However, I don't see the same conclusion. These content providers are routing around the Tier 1 providers because they're too big. Yes, it's the internet at work, routing around the inefficiencies, but not because of T1 business practices, but because they get better, cheaper service doing it themselves.

These aren't new non-Tier-1 major backbone providers. They're simply behemoths who've outgrown the playground. They're not reselling their access, they're providing bridges into the other silos. To me, this is a disheartening turn of events. While I don't see any of these companies cutting off access to the other silos (becoming AOL 2.0), they're locking up access in direct business-to-business agreements. If MS and Google decide to provide QoS on traffic X, or entirely block traffic Y, it's a matter between those two companies. Whereas, should a T1 provider do the same thing, we'd all be up in arms. Granted, The number of players makes these kinds of scenarios unlikely, but this direct linking starts to hide these kinds of concerns.