Query Jumping?

[zomlak]

I have an idea I wonder is it worth to consider!

The current scheme lets a query to be sent to a Max(TTL) away neighbors. So when we are doing a search we are always doing search aproximately on the same set of servants. Okay it may change since some servants may go and some may come but this is not the case within 10 minutes.

But what if we don't broadcast the query but instead pass it randomly to only a neighbor. Let say query is passed Max(TTL)*N nodes. And say N is greater than 3. So the query will be broadcasted at this node instead the originating. Such a node should be away from the original search spcae. So search space will be expanded with a linear growth of network traffic.

With such a scheme it is possible to find some files that are not easily found.

Okay network pressure might be a problem. But is it actually? I don't know.

What do you think of this?

[tshdos]

That is the same as just adding 1 to the TTL which as we know right now is bad.