Originally posted by jsbuser
Ok, I think you may have bought off on the hype of the C word a little too much.
At the end of the day "the Cloud" is a bunch of generic computers and storage sitting in a data centre sucking power and needing connectivity. Some of the subscription service model and provisioning changes but the underpinning doesnt.
In fact the bulk of data still lives inside commercial or government enterprises in their own facilities and before you object I'd point out that Google has cloud services, but their own commercial data is reserved inside Google for its own ends, and the same for Amazon. It runs a cloud business and the Amazon business may be a customer of its own cloud services but it is essentially enterprise data for its own use.
Much as it was when the C word was not in the common vernacular. In fact, the bulk of system in "the Cloud" are the same systems doing the same job in largely the same way as they were before. Running applications.
Yes data is growing quickly, but that has nothing to do with any intrinsic quality of "the Cloud", data storage is storage.
Reality. Tape is still the cheapest long term storage still. High capacity slower disk has eaten into that and other hierarchies such as redundant pools of SSD with n+1 replication have also found favour in certain environments or as tiers of a more complete storage hierarchy.
Now WBT have said we can meet a 10year retention. So it should be a more viable replacement for Flash in an SSD type device. If the perfomance is better as promised, and we heard low power today, then it may be that is forms part of an improved storage hierarchy in the data centre (cloud or not).
Man ... I think you're really misrepresenting the cloud here. While the underlying technology is kind of "the same" - the could still involves computers, disks and data centres - there are a lot of aspects to "the cloud" that have rightly led to the name.
You brought tape into the discussion, but tape and "the cloud" are serving vastly different purposes. Tape is used for archiving. It is not used for data that needs to be retrievable quickly at any moment. If I'm sitting on a beach in Bali with my laptop and a wireless connection and suddenly realise I need to pull up a spreadsheet that is not physically saved on my laptop, the very best case scenario with tape would be I have the ability to remotely connect to the tape storage facility and the tape is already loaded in the drive OR I have some robotic system to access and load the tape. From there, there is a SLOW process of finding the location of my particular spreadsheet file on that tape, because unlike a HDD which just rotates and moves the read heads to the correct location, the tape has to be wound through to a particular point. This process is considered more than 1,000 times slower on average than hard drive access times.
Contrast this scenario with "cloud" storage. Part of what "cloud" means for the end-user is YOU don't know where the data physically sits - and you don't care (beyond knowing it is secure). In some circumstances it might just be on a tape - but if it is, it would be because the particular data centre operating this "cloud" has worked out some algorithm for archiving files that you are very unlikely to use, and the end user accepts that sometimes, for reasons they don't fully understand, things take quite a while to load but MOSTLY it's quick. Cloud storage really is about the end-user experience and that the user is agnostic of the underlying technologies.
The capacity the cloud offers for anywhere, any time, shared access to data is a huge evolution. Even though the physical storage media might be the same, the cloud aspect changes what people might think to store. New ways of being able to access data means new ideas about the type of data we might want to store - which means more data getting stored. Once people backed up their CDs on other CDs. Then they backed them up on USB hard drives. Now you can back up your music and movies to the cloud and your kids can access your entire library on their tablets in the back seat while you take that 30 hour road trip (if you can afford the mobile data). With that capability, people start hoarding more digital data, because it is so convenient and reliable.
This has impacts on the underlying storage technologies. Growing data centres will have different demands for speed, power consumption, reliability, durability, etc. than end users grabbing USB backup drives from Officeworks. Almost no-one's going to buy a Weebit USB drive over an SSD for their Macbook because it uses less power. Some might choose it for the speed advantage ... but data centres running cloud services will be crunching the figures and if speed advantages and power cost savings balance out increased cost of the technology, it will be Hugh.
I think there are probably lots more aspects to "the cloud" that create new storage demands, but here's one final one: Multi-site redundancy. As cloud services proliferate, one thing that will become increasingly sought after will be guarantees that syncronised clones of data exist at multiple physical locations. This will benefit speed (read and write to the nearest physical storage, minimising latency, and then changes propagate to the others) and reliability (major natural disaster causes power outage at the New York data centre and power backups fail? No problem - you're seamlessly switched to getting your data from Dublin). Again, more copies of the same data means more storage needs.
(20min delay)
