On Thu, Dec 28, 2000 at 06:59:54PM -0800, Ian Clarke wrote:
I find Mr. Bad's quoting annoying, but why do you and Brandon insist on
not putting in any notice of who you are replying to at all? That is
asking for confusion (and I know we use the same MUA)...
> > I think his point is that the main advantage of having multiple nodes
> > residing behind a shield node, compared to just using the shield node
> > itself, is that there's more disk space on the total set of nodes.
>
> Well, I think that your view of this is slightly coloured by the
> clustering proposal. I don't see this as one shield node with a load of
> nodes behind it, but where nodes randomly select a shield node whenever
> they attach their address to a DataSource. Of course, nodes want to avoid
> selecting a node controlled by Darth Vader or whoever but they can either
> rely on the fact that probabilistically it is unlikely that a random node
> will, or they can confirm node authenticity through out-of-band means.
I did understand this, but the fact that the "shadow" node still eats as
much bandwidth as it provides remains.
Also, I don't think you can be quite random, since the node doing the
"shadowing" would be best off having a list of nodes that it will shadow
for - otherwise the ability to ask a node to do unlimited data transfers
without caching the data is just asking for DOS attacks.
>
> Ian.
--
'DeCSS would be fine. Where is it?'
'Here,' Montag touched his head.
'Ah,' Granger smiled and nodded.
Oskar Sandberg
[EMAIL PROTECTED]
_______________________________________________
Freenet-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/mailman/listinfo/freenet-dev