Mirror and backup is the way to go, not a waste.
The article talks about the waste that would be if each of your 100TB arrays were using only 20-50TB and the full 100TB wasn't planned to be reached before, say 5 years.

I know the agencies can decrypt anything they want to for standard SSL encryption, and probably will if you're red-flagged somehow, but if I understand this article correctly, they can't make legal use of this information without a formal NSL letter first, right?

In that case, if I am using, say https Gmail, can the ISP technically answer such a NSL letter?
The Google IP address is visible in the TCP/IP headers, but isn't the sender and recipient name and IP hidden behind SSL?
Or are the senders and recipients visible because Gmail is using POP4 or whatever protocol that the ISP can read?

$30/GB/month, now I know you work at Microsoft! ^-^
I think this was about the same outrageous cost of storage in the corporate datacenter farm 2 years ago, although there were also different cheaper options for larger storage needs. And there was also a one-time setup fee before even the monthly bills...

So I made a proposal for our own department for a 8TB storage solution, consolidating several smaller file servers into one (plus another for the backup) with this DAS enclosure: http://www.pc-pitstop.com/sata_enclosures/scsat84xt.asp
Cost was about $6K for the 2 enclosures, 2 RAID SATA controllers and 18 1TB Adaptec drives. The servers were scavenged, the OS and backup software was for free, the set up was also provided by us for free. This was a one-time cost for less than a week of what the IT department would bill us, and zero hassle compared to all the paperwork required to fill in such a request, which basically forces you to overestimate the storage space need, as you don't want to repeat this horrific experience. Maintenance was supposed to be minimal, the disks were configured in RAID 6 to allow for 2 simultaneous drive failures, we had also planned for a couple of spare disks already mounted in spare trays to quickly hot-swap failed ones. Basically maintenance was required only to restore the backups or specific files to an earlier restore point in case of data corruption.

You might get a problem with corporate policies that require the use of the centralized datacenter. I left that department shortly after my proposal, don't think it went anywhere.
I'd make the same proposal today, except that, if the storage need requires it, I'd probably get 16-drive DAS towers or rackmounts with 2TB disks (just 8 disks is a waste for RAID 6), about $6K for 28TB of usable space out of 32TB raw space, plus another 28TB for the backup server set in a separate location.

It's quite bad that the $1M / 100TB cost is not detailed in any way in this article, it makes all attempts of comparison futile and impossible.
Thankfully several commenters provided their input, some even mentioning billing $5M per 100TB to their lawyers customers.

My guess is that this $1M represents partially the costs of the hardware/software configurations in existing datacenters, with most of these possibly purchased 2 years ago.
2 years is a pretty long time in IT, with many technological and financial changes, such as SATA III, 2TB disks and SSD, the latter still being rather immature and expensive.

Given the relatively low cost of the hardware, it does make sense to implement a better aggregation/allocation infrastructure, and disconnect the unused servers, but keep them on hold to add extra storage in a moment's notice, or to scavenge spare hard disks. They won't cost a dime if they are plugged but powered off, it's the online storage and maintenance that costs $1M/100TB (if that much.)

That's not how I read the article, it describes datacenters with 40-60% of unused space, I think the goal if to reduce this to something like 20%.
If you manage 10PB at 50% usage, you can reduce it to 6PB at 20% usage, thus saving $40M (at $1M/100TB.)

Hopefully SSD disks will relieve RAM, once the market becomes mature and cheaper at even higher densities.
$3000 for a 1TB SATA II drive is a little bit excessive... ^-^

The article's focus is more about datacenters reducing their unused space from say 40% to 20% for petabytes of storage.
You're dead if you let unused space down to 5%, let alone 1%. Even home users get red flags when their disks go below 10% of free space.

You can't rely on the OS and extra storage space to fully restore deleted files, the OS can reallocate that space at any time.
It's just pure luck each time if you can, although I agree extra space increases your chances.

You should rely on your backups and maybe custom scripts to trap all file delete requests at low level instead, but I don't even know if that's feasible. That would be totally rad!

60 watts? A 2TB performance SATA III drive (not the "green" low power drives) is about 8W (6 on idle, 9 on load).
So that's 8x24*365 = 70KW, more like $6.
http://www.seagate.com/www/en-us/products/internal-storage/barracuda-xt-kit/#tTabContentSpecifications

Anyway, the cost of the hardware is almost zero compared to the other costs, except is you use SAS or SSD drives.

Adding more servers is wasteful only if you have poor storage management.
If you need more space, you should be able to just add a new server and allocate that extra space to the customers who need it, or not allocate anything at all and bill the customers for the actual space/transfer used or for the extra TBs above their quota. You probably don't even need quotas at all with a smart storage management software.

I used that backblaze analogy too, but after yours, @ano:
http://hardware.slashdot.org/comments.pl?sid=1735418&cid=33075574

It comes up to $45K/90TB tops, in just one 4U enclosure, management overhead and real estate included. ($50K/100TB)
To answer @spazimodo, the Taiwan-built Barracuda XT disks have low failure rates compared to the Seagate disks built in China, and you can combine them in RAID 6 or 60 with Adaptec controllers for $1-2K more. Why would you ever use RAID 5 anyway, that's insane. We're talking raw storage here, so backup hardware, snapshots, replication are exactly the same as the 1M/100TB raw space estimate. If you want real usable space data all costs included, get 2 such servers, let's say at $50K each with RAID 6 controllers and fiber optics cluster connection. That's 74TB of usable space, plus another 74TB on the backup clustered server, or $135K/100TB usable space.
Even a backblaze enclosure comprised exclusively of SSD disks for performance would cost only $72K per enclosure, hardware and 1st year costs included, about $640K/100TB. But I assume this level of performance is far above the $1M figure, assumedly for SAS disks, and which also probably spreads out the hardware cost onto 3 years, so it's more like $0.5M for pure SSD 100TB. It would really help if the article would detail that $1M cost.

SSD costs assuming 256GB disks @ $700 each. 1 backblaze 4U server would provide only 11TB of SSD raw space, 9TB of usable RAID 6 space.

Well 10TB is rather small and can easily be filled up, so you need the low usage rate, but for datacenters with petabytes of storage space, which are what the article is mostly about, you can aggregate the unused space, remove some, and increase the global usage rate quite a lot, while still allowing for huge increase of storage usage from many of your customers or departments. Compared to the whole capacity of the datacenter, these increases are rather diluted to small percentages.

Not really. Google designs and builds their own servers.
The "super expensive storage solutions" are for suckers.
http://news.cnet.com/8301-1001_3-10209580-92.html

These expensive solutions are probably the reason why the analyst mentions saving $1M for each 100TB removed.
With 4U enclosures like backblaze's, you get 90TB for $11K of hardware and $6K (45 disks @ 8WH) of power usage per year.
An IT operator can control dozens of such enclosures, let's say a conservative 2 dozens. So $160K salary / 24 enclosures is $7K.
Add $7K for a full time dev and custom storage management software, and $14K for management (still for 24 enclosures).
That's still about $45K for 90TB all included, exactly 20 times less than the mentioned $1M for 100TB.

http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/
I replaced the 1.5TB disks with Seagate Barracuda XT SATA 6Gb/s 2TB disks at $200 on newegg in this computation.
Seagate's other models built in China have lots of problems that the XT doesn't seem to have.

A new Windows rewritten from scratch would really help a whole new class of efficient and reliable embedded and consumer systems, maybe an OS that would make multiplatform programming easier too (developing apps simultaneously for Windows, Mac OS and Linux.)

But that would be a tremendous job, you'd basically need to rewrite from scratch all the applications and especially the developers' tools. Even then devs would need a lot of training about the changes and new features offered by the new OS, how to port their existing apps, etc. All software companies would need to adapt too. Not too sure we'll find many supporters for this cause.

*you => read @Liquidrage