"The affordability and ever increasing capacity of data storage is affecting all aspects of business and daily life. Peter Cochrane predicts what's coming next and what sort of changes will occur once the petabyte is commonplace."
Peter Cochrane has some wonderful observations:
"...TB storage systems anywhere. It seems that Moore's Law has migrated to storage. We are now within an ace of the TB PC and laptop..."
"Only four years ago my son built and installed our first terabyte (TB), or about 1,000GB, home storage server at a cost of $4,000. A month ago we installed the second TB for only $600. What a drop in price and physical size - two parallel drives instead of five."
"We are now within an ace of the TB PC and laptop. My guess is that by 2015 we will also be using around 10GB of RAM. You may think this ought to be enough for most individuals and businesses but when you imagine movies may soon be distributed digitally at around 6 to 12GB each, you can see it all being eaten away. And how about all 26 million books in the Library of Congress? It will need around 1TB too.
The technology to realise hard drives up to and beyond 100TB is emerging in the R&D labs whilst petabyte (PB), or about 1,000TB, systems have already been engineered at great expense for specialised applications. I see no limit on the horizon to the growth of information storage as we have yet to enter the quantum world of the really small."
"But I think there is a more fundamental question we should be asking. We will be faced with an infinite and distributed sea of bits - much of which will not be itemised, catalogued, categorised or labelled in any way. Moreover, there will be duplication and storage corruption on a global scale. The really big question is: how the heck are we going to identify and locate anything in this mostly uncharted sea?
Will today's search techniques fit the bill? I think not - they are far too crude. We are going to need much higher degrees of sophistication to the point of machine cognition in recognising scenes, contexts and relationships in our data. The first machine capable of realising such a capability will probably arrive around 2010 in the form of an adaptive supercomputer and, given our rate of progress, on a PC scale by 2030 or earlier.
Now for the kicker: I cannot see how a fully centralised or fully distributed system spanning the global network will satisfy our future needs. Seems to me an agent-based hybrid scheme is the only contender that will do. Right now I don't see how we are going to do that but artificial life plus artificial intelligence in an evolvable form seem to be strong contenders."
The Other Exponentials - Technology Review
Lifetime Digital Memory
What 2034 will bring
Moore's Lore - Corante
John Battelle's Searchblog
Search Engine Watch