Apple Rumored to Purchase 12 Petabytes of Storage for iTunes

[athleticswimmer]

A Yottabyte is nothing. Googol, now that is something! Googol is the number one followed by one hundred zeros.

Googol: 10,000,000,000,000,000,000,000,000,000,000,000,000 ,000,000,000,000,000,000,000,000,000,000,000,000,0 00,000,000,000,000,000,000,000,000

But there is something even bigger called a Googolplex. The number is so large I cant even write it. its Googol * 10!!!! So it's 10 times the size of Googol THATS CRAZY!!!! A Googolplex laughs at a Googol

*But since were talking about storage and not numbers. Im going to call it a Googolbyte......

[feidhlim1986]

A Yottabyte is nothing. Googol, now that is something! Googol is the number one followed by one hundred zeros.

Googol: 10,000,000,000,000,000,000,000,000,000,000,000,000 ,000,000,000,000,000,000,000,000,000,000,000,000,0 00,000,000,000,000,000,000,000,000

But there is something even bigger called a Googolplex. The number is so large I cant even write it. its Googol * 10!!!! So it's 10 times the size of Googol THATS CRAZY!!!! A Googolplex laughs at a Googol

*But since were talking about storage and not numbers. Im going to call it a Googolbyte......

What about a Zillion or even better a Gajillion

[mvhurlburt]

I'm sure they won't need quite as much redundancy as you're thinking. If 8 million users all have the same Ke$ha song in their cloud "locker". They won't need 8 million copies of that song. They'll only need a few (I have no clue exactly how many, but the point is illustrated). There's going to be so much cross-over, that they could likely get even more than 12 effective petebytes. Think about it. An album that's 100mb could be "stored" by 10 million people and still not take up more than 100mb. I realize they would still need multiple copies, but not 10 million copies.

P.S. - I'm pretty sure that Dropbox already does this. I uploaded a file one time (I can't remember what it was, but it was large, and it was a common file. i.e. - something that other users would have already had in their Drobboxes), and while it should have taken 10-15 minutes to upload, I noticed that Dropbox already said that it was all synced up. I logged into the website and sure enough, my file was already there and accessible. Why would Dropbox waste the bandwidth and drive space to store another copy of a file that they already have? They can check file sizes or whatever to make sure that the file is indeed identical and not just the same file name.

I agree. This is very likely the way apple will deal with redundancy. All iTunes store content that a given user has will already be present and most likely be backed up in the form of a list or database, etc. For other outside or unique content one of two things will likely occur:

Let's say the content was ripped from a CD, DVD, etc. it will likely be compared with what the iTunes store already has on file. 1 - if the file is identical to what is already on file with iTunes an entry would be added to that users list of content

2 - If that particular song, movie, etc. is present in iTunes but the user's version is not identical or that file does not exist in iTunes at all it will be backed up to Apple's servers and will likely exist as two copies for the sake of redundancy.

To further the efficiency of storage space the unique file in example two could also be compared against all other unique files with the same file name and size though that would likely be an inefficient use of CPU cycles and could be time consuming.

Either way, whether apple chooses to apply a similar model to what I have described above, they certainly will not have 8 million copies of the same data. A majority will be based off of a central iTunes content database. How they handle unique data may vary though. It's certainly fair to say that actually copying EVERY users media to the cloud would be EXTREMELY inefficient and downright stupid. The actual storage space available after redundancy will likely be 5-8PB out of the original 12PB depending on the redundancy scheme Apple may choose. It's gonna be one big as$ RAID array though.

P.S. - It is important to consider that we don't know what Apple currently has on hand in terms of storage space. This 12PB is in addition to. Also they have been working on this behemoth of a datacenter for years now so it's kind of hard to really say what they may or may not already have.

[athleticswimmer]

Googolplexian is the largest number. yes larger than googolplex. To see how many zeros visit this page: Googolplexian: The world's largest number with a name. Worlds biggest. and it doesnt even show you all the zeros lol!!! ok back on topic every1.

[politicalslug]

I'm sure they won't need quite as much redundancy as you're thinking. If 8 million users all have the same Ke$ha song in their cloud "locker". They won't need 8 million copies of that song. They'll only need a few (I have no clue exactly how many, but the point is illustrated). There's going to be so much cross-over, that they could likely get even more than 12 effective petebytes. Think about it. An album that's 100mb could be "stored" by 10 million people and still not take up more than 100mb. I realize they would still need multiple copies, but not 10 million copies.

P.S. - I'm pretty sure that Dropbox already does this. I uploaded a file one time (I can't remember what it was, but it was large, and it was a common file. i.e. - something that other users would have already had in their Drobboxes), and while it should have taken 10-15 minutes to upload, I noticed that Dropbox already said that it was all synced up. I logged into the website and sure enough, my file was already there and accessible. Why would Dropbox waste the bandwidth and drive space to store another copy of a file that they already have? They can check file sizes or whatever to make sure that the file is indeed identical and not just the same file name.

Logically what you're saying makes sense, but you forget the most important thing: licensing. Whether Apple can treat the files as music, thus maintaining one file instead of 100 duplicates, then they would not need nearly that much storage (assuming people only store music, forgoing pictures and video). If, on the other hand (as has been suggested on this site and others) Apple must treat the files as arbitrary user data then they would have to keep separate files in order to satisfy the record labels. This proposed cloud system will not exist in a bubble. Apple is constantly under scrutiny by the labels. Music aside, if users are allowed to upload HD video from their iPhones, iPods, and iPads, then this storage for video alone won't be sufficient (assuming Apple doesn't extremely overcompress video as YouTube has chosen to do). And picture libraries are huge (mine is 20 GB and I haven't taken a picture in two years).

[mlauth]

I wouldn't get too excited in terms of thinking tons of content will be stored on 12PB.
When you take the same files stored multiple times, data backup and redundancy into account, it's more realistic to look at it as more like 1 or 2 PB of media stored across the 12PB

Its called data de-duplication. Why back up 10,000 of the exact same file when you can back it up once.

How much does this amount of storage cost? I took a quick look on the Isilon website but couldn't find anything.

A rough estimate:

I just got a quote from EMC for 25TB of 15K SAS 1 shelf of SSD and 1 shelf of 10k SAS.

My quote:
747,945$

I would bet it is in the 100s of millions.

[feidhlim1986]

Its called data de-duplication. Why back up 10,000 of the exact same file when you can back it up once.

Good luck having 10,000 users accessing two files

[carlito9505]

Skip this place go back to yours
Our lips first touch outside your doors


Anyway back on topic here um theres roughly 1,000,000 GB in 1 PB right?

Well, I know there are 1 quadrillion bytes in one Petabyte

And 1 Septillion bytes in one Yotta byte

I wonder if anyone has tredecillion (one set of three zeros less then googleplex) of anything?!?!?

[Megaorange]

skip this place go back to yours
our lips first touch outside your doors


anyway back on topic here um theres roughly 1,000,000 gb in 1 pb right?

1048576 gb = 1 pb

[stevelucky]

Logically what you're saying makes sense, but you forget the most important thing: licensing. Whether Apple can treat the files as music, thus maintaining one file instead of 100 duplicates, then they would not need nearly that much storage (assuming people only store music, forgoing pictures and video). If, on the other hand (as has been suggested on this site and others) Apple must treat the files as arbitrary user data then they would have to keep separate files in order to satisfy the record labels. This proposed cloud system will not exist in a bubble. Apple is constantly under scrutiny by the labels. Music aside, if users are allowed to upload HD video from their iPhones, iPods, and iPads, then this storage for video alone won't be sufficient (assuming Apple doesn't extremely overcompress video as YouTube has chosen to do). And picture libraries are huge (mine is 20 GB and I haven't taken a picture in two years).

I don't think that's accurate. Licensing is only a concern with how the music is initially distributed, I.e. - how the user gets the music in the first place, not where it's stored. How many people store their music on Dropbox, yet Dropbox has no licensing concerns. Who's to say that I didn't rip a physical CD and store it in this fabled iTunes cloud? How would Apple demonstrate appropriate licensing? They couldn't. I think licensing concerns would remain just as it has: in distribution, not storage.

Though I do agree, factoring video definitely poses a challenge in and of itself. I honestly picture this cloud (if it exists) as mainly for music. That's what the majority of people would be interested in.

Though a part of me would rather picture this cloud more like the cloud city in Star Wars. I think that's something we could all get behind.

[alexevo]

I bought a spindle of CD-Rs today. Where is my article?

I don't think that's accurate. Licensing is only a concern with how the music is initially distributed, I.e. - how the user gets the music in the first place, not where it's stored. How many people store their music on Dropbox, yet Dropbox has no licensing concerns. Who's to say that I didn't rip a physical CD and store it in this fabled iTunes cloud? How would Apple demonstrate appropriate licensing? They couldn't. I think licensing concerns would remain just as it has: in distribution, not storage.

Though I do agree, factoring video definitely poses a challenge in and of itself. I honestly picture this cloud (if it exists) as mainly for music. That's what the majority of people would be interested in.

Though a part of me would rather picture this cloud more like the cloud city in Star Wars. I think that's something we could all get behind.

Well said if I do say so myself. Full marks lad!