60% of bitcoin traffic
When you say 3 isps do you mean the big tier one transit providers?
Those networks that probably see a large fraction of any international traffic
Attacks on Bitcoin just keep coming: ETH Zurich boffins have worked with Aviv Zohar of The Hebrew University in Israel to show off how to attack the crypto-currency via the Internet's routing infrastructure. That's problematic for Bitcoin's developers, because they don't control the attack vector, the venerable Border Gateway …
... although it is certainly the most popular blockchain implementation these days. However, it is going to fade away at some point as it can't support the volumes and latency requirements needed for global mass payments systems (until this becomes general concensus, Bitcoin owners are in for some significant volatility risks, while speculators will have a lot of thrill, fun and the opportunity of their lifetime ...).
Many other blockchains for a lot of different purposes are under development right now, chances are that those might also be affected by vulnerabilities in the Internet routing Infrastructure.
This is basically a restatement of CAP: you can't have a distributed system which is Consistent, Available for writes, and Partition Tolerant at the same time.
As the blockchain doesn't have a fixed set of authoritative nodes, it's very hard to prevent updates in the case of partition.
In a normal system you could design for AP with eventual consistency, by merging together the changes. However in the bitcoin case you don't necessarily trust a small subset of nodes to make independent updates to the history. And in any case, double-spending could take place during the split-brain period, which can't possibly be merged.
"I mean, what country is going to charge someone for stealing BC anyways lol Bitlanders and the Bitpolice, in their bitmobile, and put me in bitcell hahahaha"
For starters, the USA:
http://www.bbc.co.uk/news/technology-34114483
Whilst HMRC acknowledges that VAT is not due on bitcoin mining and similar activities, they have pointed out that capital gain, income and corporation tax rules do apply. This would suggest that a bitcoin thief could not assume that the UK authorities would be uninterested.
The whole point of the blockchain is that you're hoping the good guys work faster than the bad guys. If a nation state dedicated all their resources to bitcoin mining, they could, in theory, out-calculate a lot of bit-coin transaction, effectively overtaking the valid blockchain with their own version they provide.
Now, you're saying you don't even need to out-muscle them, if you control the gateway.
So the blockchain is only secure if there are enough people, spread throughout the world, all on the side of good out-computing evil entities. How long will this be the case? My understanding is that it's not profitable for any but the most dedicated users to run their own mining farms these days. The game is changing from every user running their PCs overnight to massive compute farms in isolated areas with cheap electricity and fast internet (Scandinavia for example). Doesn't seem like Bitcoin is moving in a stable direction.
This does make Bitcoin and their blockchain method look a bit shaky. There are ways of improving it, but not without some kind of protocol overhaul that creates a bi-direction, ring distribution for the most trusted nodes (node X can't reach Y, but can reroute, or ask node Z which can reach Y in proxy of X), but that is a gross simplification. Spending money to make it work through a CDN, and the accompanying rework to the protocol for that endeavor is possible. Those are two off the top off my tiny head. Still, that does pose a problem to the leaf nodes unable to dual route their BC traffic with the main dist ring, and conceivably multiple rings stemming from that one. Recovering from disconnection is one thing, but with enough bandwidth and multiple peering connections could make verifying the chain possible when a direct connection/update/verification is not immediately possible. There could also be some kind of fix to the problems cause by disconnection delays, but that is too deep for me to wade in with my limited knowledge of Bitcoin. Very interesting to watch this new currency develop nonetheless.
Any good currency, or language needs an army to back it up. The world's main currencies seem to have this made (I could question the euro, but there are a couple of armies there).
Three cases:
Esperanto: Fail, no army!
Ebonics: Fail, no army!
Bitcoin: Bound to fail, no army. May exist as a "stock" and traded as such.
And so it goes...