« It's art, stupid | Main | My backup requirements... »
June 29, 2006
The Next Big Thing?
I think the NEXT BIG THING will be boring. It won't be search, it won't be social, it won't be tagging and folksonomies - although it might well encompass all these things.
The Next Big Thing will be storage. Specifically remote, accessible, easily synched backup services.
For your digital camera you want to have the equivalent of the drawer where you store all the family photos. For your music you want the equivalent of a cardboard box full of gooved (LP's) and pitted (CD's) vinyl where you can store all your MP3's.
If you are a developer/designer or business person you need to store your email, documents, code, images and versions of all these.
The Next Big Thing will allow you to easily do all of these things, easily, simply, from any computer at any time.
Here follows a brain-dump think through (more for my own edification than anything else - ignore if you will):
I have thought for a long time that a SET@home-bittorrent distributed application would work well for this.
The idea would be to have an application running on each contributors computer that sets aside a little bit of space on the hard drive (say 1% of the toal available). The application allows easy addition, retrieval, synching and even deletion of stored data.
The data is submitted to the distributed-p2p network where it is broken into pieces bittorrent style before having pieces distributed to all the members of the network where it is stored on their PC's.
Immediately you can see a problem with this - imagine the network has 1 million people on it and all those 1 million people have on average 2GB of storage reserved for the network. Imagine now that all the members wanted to store 1GB of information. The system would have to be storing multiple copies of your data to ensure data integrity so lets say at a very low estimate it stores three copies of the data you send it. Suddenly the system is overloaded.
Even coming close to overload the system would have problems. The point is to have yur data distributed all over the world so that it is safe but approaching overload the system would start clumping large numbers of bits of your files together. This reduces the security of the data and renders the system a little bit pointless.
In reality most people would be storing a hell of a lot more information that that. Photo and MP3 collections for Joe Soap currently could easily run to 30+Gigs.
Currently we have been talking about a peer-based service. But what if we could build a business that hosts servers that also contribute to the storage capacity? Big boxes that run the client and share out 100% of their drives as storage. That would definitely help the situation. This is not practical either. The machines would need to be stored in different places all over the planet, and there would need to be millions of them - impractical for a start-up.
Possibly you could get each hosting provider to dedicate a rack of machines to the client/storage. What incentives would there be? Money? Doubtful, the users could be tiered as so:
- Basic - store up to 20GB - retrieve whenever - refresh once per week - includes data integrity only
- Heavy - store up to 50GB - retrieve whenever - refresh daily - includes data integrity only
- Developer Basic - store up to 250GB - retrieve whener - refresh whenever - data integrity and versioning
- Developer Heavy - store up to 1TB - retrieve whenever - refresh whenever - data integritu and versioning
You could also add storage increases on top of each package ie. by the percentage of current level - so a 10% increase costs roughly the same whatever package you have.
In conclusion - boil the sea? - I think so.. pity :(
Posted by dottie at June 29, 2006 6:56 PM