[norea-armozel]

You're confusing the issue by focusing on protocols versus actual physical implementations (data centers, trunk lines, etc). The physical installations for what we call the Internet are centralized. Companies like Level 3 might put some redundancy but at some point the cost of redundancy out weighs its benefits for them and other companies like them. This is especially true of consumer financial services like banking. If an attacker wanted to disrupt the United States they only have to do it to banking to cause a panic. They could easily ignore emergency services, hospitals, and even the government itself (outside of ACH) while doing this. And it would be such a mess that we couldn't resolve it immediately. The happiest outcome is the disruption is only for a few hours but the more likely outcome is possibly days or weeks of disruption where a large part of the banking system would be inoperable. It doesn't matter if you used TCP/IP or switch based communications the outcome is the same: the American economy shaken and possibly worse. So, we can take all day about Facebook and Diaspora but neither of those services do anything important for the average user like your bank which also uses the same centralized infrastructure. There is no Diaspora for banking and not one that's widely used or not using the current banking/financial transfer systems which are centralized.