On 10/04/14 16:46, Robert Hailey wrote: > On 2014/04/10 (Apr), at 7:21 AM, Matthew Toseland wrote: > >> Thoughts? Is it still worthwhile to do only-route-high-HTL-to-core-nodes >> first... > I'm still trying to understand the full effect of this... to me, low-htl > routing (deep/tail-end) is a more interesting (e.g. for performance & > actually finding the data), but I can certainly appreciate wanting to shield > ourselves from newbie nodes. IMHO a large proportion of requests find the data at high HTL. About 35% according to my stats. >> ... or do we need to go straight to tunnels? > I've always been a fan of tunnels, or... "union routing done right" (as there > are several ways to do it), but some part of me thinks we need to fix our > current problems before we (potentially) layer on more. Union routing done right? >> And if we have tunnels, >> do we still need only-route-high-HTL-to-core-nodes or would it be better >> to rely solely on tunnels? > At the face of it, I would presume it would be better to *avoid* establishing > tunnels through new & untrusted nodes... unless all requests go through > tunnels, in which case it's about as sensitive as the requests you would be > protecting (high htl, inserts, etc). So it seems the ideas are related, as > there is probably a better chance and danger of dropping into a Sybil network > or MAST node if they are "new", right? Right, but the main reason is for performance: We don't want to tunnel through unreliable, slow nodes. Which is a security issue too as when the tunnels break we have to make more.
Doesn't answer the question - do we still need "route high htl requests only to core nodes" if we have tunnels?
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Devl mailing list Devl@freenetproject.org https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
