I have been following these Bitcoin stories with some interest and I have a technical question maybe Slashdot can answer. It seems to me that Bitcoin has problems with scalability.
My impression is that to perform a Bitcoin transaction you have to download the entire history of Bitcoin transactions before you can get started and the entire network has to confirm your transaction and it's authenticity to prevent double-spending of coins. So far, so good. The problem is that this is an all-to-all network in terms of storage and processing requirements and its needs seem to scale exponentially. Last year, I hear there was 2 GB of storage required and this year 6 GB.
My question is: doesn't that imply that Bitcoin won't scale and will eventually fizzle out due to impractical storage and processing requirements? Who will want to download 100 GB or 1 TB of history of transactions of the entire world just to buy a sandwich?
And doesn't that imply that there is room for Bitcoin 2.0, a new virtual currency with better scalability and possibly other improved characteristics? Compare Bitcoin with the Internet itself. The Internet is far more scalable because it is broken down into subnets and features more point-to-point communication and no need of transaction history which makes it far more scalable.
Hypothetically, couldn't another virtual currency like Bitcoin be devised, that separates its users into subcommunities based on their frequency of transaction or physical location, and thus feature more efficient local transactions, which most transactions are, while allowing the occasional transaction to be routed between subcommunities? It could also feature a shorter history that only remembers a certain period worth of transactions, like one month, along with everyone's balance, thus avoiding the need of storing the entire history. Further, it could develop a reputation system that enhances the ability of the system to reject bogus transaction.
tl, dr: How is Bitcoin supposed to scale to more usage given its exponentially increasing storage requirements, and what can we replace it with that is more scalable, being better structured, with shorter history, or more efficiently rejecting bogus transactions?