[thomaslangston]

> It's hard enough to secure centralized systems using proprietary code behind firewalls managed by an army of experts. How can a decentralized system possibly avoid becoming a 0day distributed botnet?

I do not understand your concern.

Centralized systems are easier to crack, not harder. Proprietary code may or may not be easier to crack, but I'd lean towards easier. Distributed systems do not require you to not have firewalls. Distributed systems do not require you to not have experts managing them. Nodes of a distributed system do not have to obey the commands of other nodes or have a central control mechanism.

How would a decentralized system ever become a 0day botnet?

[pacala]

The point of Freedombox is that everyone has his data in his own home. If average Joe wants to share this data with his friends, average Joe is going to run a server in his home. If the server code has a 0day exploit, average Joe server easily becomes a node in a botnet.

How do you address the problem of managing a fleet of 800 Million servers running in 800 Million homes? Who are the experts that manage this problem and who finances their work? Why would average Joe trust them with his data? What is the technical device through which they have admin access to 800 Million servers in 800 Million homes?

[thomaslangston]

That may be the setup of that project, but it isn't the status of every decentralized system. A decentralized system could as easily be funded by individual users buying hosted space with professional admins.

I don't see the 0day exploit window of one additional application running on local users' boxes to be much cause for alarm. Somehow we muddled through decades of Windows boxes with everyone joining the Borg.