[realcr]

It depends. If you own all the computers, you could just pick random identifiers (Of some large space). You could also choose the identifiers to be a public key, or a hash on a public key. Then when talking to a node claiming that he has a certain ID, you could ask him to prove it.

Even when using the public key as an ID idea, it is still possible to get IDs as close as one wants to a particular number. (It is a bit harder computationally, but still possible).

There are more secure things you can do. Shortly, you would prefer the network itself to choose the ID for a new node, and not let the new node choose it itself. It appears that even this will not give you a secure enough solution. I will write about it in future articles.

[newlog]

hey!

First of all, awesome explanation, thanks!

What I didn't understand from your explanation (I don't know if it's explained or if I missed it) is how you identify which node has the value for the key K when the keys are not in the Bs domain. How do you know where to stop the search? What if you have gone to a further node?

I'd be very grateful if you can clarify that :)