View Full Version : New hard drive?

01-24-2012, 07:54 PM

Researchers at IBMís Almaden researcher lab in California are working on a 120 petabyte hard disk array. For those of you playing along at home, thatís 120 million gigabytes spread out over 200,000 hard drives. The array is being assembled for an as-of-yet unnamed client, and will be used to run complex modeling simulations, like those used in meteorology.

Hereís some stats to drool over: With 120 petabytes of storage size, Technology Review says that the array could store a trillion files, 24 billion 5 megabyte MP3s, or 60 copies of the 150 billion page Internet Archive WayBack Machine. The largest arrays currently available are around 15 petabytes, about a tenth the size of the IBM array. Bruce Hillsberg, the lead on the project, must be very proud of his water-cooled memory monstrosity.

Beyond its voluminous size, the IBM array has a few software tricks up its sleeve. One of these, called the General Parallel File System or GPFS allows of super fast indexing of massive amounts of information. With GPFS, the computer attached to the array can also read and write different parts of individual files at the same time, despite their being split up across the myriad of drives. Technology Review writes that a recent GPFS demonstration indexed 10 billion files in 43 minutes, which shattered the previous record of one billion files in three hours.

The massive disk array also has a unique system that accounts for the inevitable death of hardisks. From Technology Review:

When a lone disk dies, the system pulls data from other drives and writes it to the diskís replacement slowly, so the supercomputer can continue working. If more failures occur among nearby drives, the rebuilding process speeds up to avoid the possibility that yet another failure occurs and wipes out some data permanently. Hillsberg says that the result is a system that should not lose any data for a million years without making any compromises on performance.

While currently ludicrous and over-the-top, Hillsberg points out that drives like this are sure to become commonplace as the world moves toward a more cloud-based computing environment. Cloud networks demand high capacity storage, high reliability (after all, why would you trust your precious family photos to a system that canít guarantee their safety?), and fast speeds so as to seem natural for users. IBMís 120 petabyte array is a monster, to be sure, but itís sure to soon be eclipsed by more and greater machines.


01-25-2012, 06:30 AM
There's where the NSA is gonna store it's take from the next generation of predator. Yes, I'm serious and that's about the only big money application that matches that type of processing and capacity.

01-25-2012, 08:29 PM
or there going to store the data for this


Next 5 In 5: IBM Predicts Mind-Reading Computers Of The Future (VIDEO)

The Huffington Post Ramona Emerson First Posted: 12/20/11 06:22 PM ET Updated: 12/20/11 06:22 PM ET
share this story
Smarter Ideas , The Future , Ibm , Video , 5 In 5 Ibm , Ibm 5 In 5 , Ibm 5 In 5 Predictions , Ibm Next 5 In 5 , Ibm Next 5 In 5 Forecast , Mind Reading Computers , Next 5 In 5 , Next 5 In 5 Forecast , Next 5 In 5 Ibm , Next 5 In 5 Predictions , Predictions Next 5 In 5 , Technology News

Forget Siri. According to IBM, in five years computers will be able to do a lot more than listen to your voice: They'll be able to read your mind.

IBM has announced its "Next 5 In 5 Forecast," an annual list of five predictions about the technologies the company's innovators think we'll see in the next five years.

The Forecast, which has been around since 2006, has had some misses in the past, such as the 2007 prediction that doctors would develop super-senses to smell illness; but on the whole, the lists have been fairly prescient. A more on-the mark prediction in 2007 said that cellphones would soon take the place of wallets, banks and concierges. Considering the ubiquity of phone payment services, online banking and review apps, IBM was right on. When the folks at IBM predicted that "You will talk to the Web and the Web will talk back" in 2008, they probably didn't know it would only be three years until that became a mainstream reality with the introduction of Apple's voice-activated personal assistant Siri.

One of this year's coolest forecasts -- that computers will soon be able to read your mind -- is already sort of a reality. For example, scientists at the University of Berkeley are able to use images of brain activity to roughly reproduce the picture or video that a person was watching when the activity occurred. Similarly, the EPOC Neuroheadset from electronics company Emotiv uses sensors mounted on the scalp to allow people with neurological disorders, such as locked-in syndrome, to use their minds to move objects on a computer screen.

The people at IBM imagine that even more advanced brain-computer interfaces could bring this kind of interaction to the masses. In a few years, with the help of sensors connected to their mobile phones, people may be able to make a call just by thinking about making a call. This could be pretty awesome, though it might prove embarrassing for the more obsessive among us.

Kevin Brown, a member of IBM's Emerging Technology Services team, outlines other potential uses of this technology in a post on the IBM Research Blog. According to Brown, if people's thoughts were automatically uploaded to a central computer, a heat map could be created showing how people in different areas of a city were feeling. Creative fields, he suggests, may offer the most interesting use of this potential technology. He writes that musicians could use the mind-reading devices to compose music based directly on their thoughts. The ability to "upload" a story or painting directly from your brain would likely be of interest to anyone who has dealt with the exasperating chasm that exists between idea and implementation.

Check out more about this and IBM's other Next 5 In 5 predictions in our slideshow below. To see more of what the future might hold, take a look at 11 mind-blowing concepts that could be right around the corner.

naturally if they announce it they're probably 10-15 years behind whats still secret

01-26-2012, 01:32 AM
1. Requires a hours long MRI of every single person (http://newscenter.berkeley.edu/2011/09/22/brain-movies/) you want to read. Even if the government had some magical way of doing it when we went through airport scanners or the doors of our DMV, there would be no good reason to have it be this scale. You'd need to distribute thousands upon millions of the scanners, have dark fiber to push the results back to process and involve so many thousand people that there would be some sort of "tell". Area 51 was easy to find because of the number of government employees working there. There also were public signs it was a test ground for aerospace to to observable phenomena. To boot: we live in an age of cell phone cameras, WikiLeaks, TOR and Anonymous. If the government was trying to work on some Manhattan Project scale directive, it wouldn't STAY secret. Hell, the Manhattan project wasn't even THAT secret in it's time. People look for areas on Google Maps that aren't refreshed regularly and track them. They built secret cities, compartmentalized information and did things that scientists of diverse political beliefs just simply wouldn't put up with in the information age.

2. That's the wrong kind of math. Weather models are are solving partial differential equations, then using statistical techniques based on observations and the laws of physics to remove clutter and give you indications of what is most likely to happen. These are not simple problems, though they are repeated a lot of times and then "hashed" to give likely outcomes. It requires more complex cores to do parmetric math that you can't logistically brute force. The brain scan decoder is a multi-step process that involves a lot of comparison and matrix transformation. Cell processors (like video cards) are better at that work. Similar to the hashes used in BitCoin (http://bitcoin.org/). One application that is like weather that "makes sense" is keyword searches directed to find terrorists or tax evaders or whatever the fuck the feds want to prosecute. Our current AVR software is either "broad base" that looks for one word across a wide range of speakers (which requires complex math to draw the line to give a yes/no answer) or trained that adapts itself to one speaker and is highly accurate with a big vocabulary. There's interest and the ability to load a pattern generated with some lead processing time into the powerful decoders and then grind through the database of billions of conversations that the feds would store to look for timelines on. Another good one is trying to create rainbow tables to bust through PGP and other similar commercial and government used encryption algorithms. The actual tables wouldn't be that hard to work with once they were produced, but the pools to create them would involve insane amounts of storage and processing power. It could also be a billionare trying to jump start SETI by saving every piece of radio chatter we intake or scientologists trying to do something L. Ron told em to.

02-06-2012, 12:43 AM
I think they forgot to mention it will take 1.21 gigawatts to power up and cool all those spindles....