Last year I came up with a list of things that Freenet needs to make a real
impact, and concluded that most of them were impossible. However, we have made
considerable progress on many of them, at least in theory; we know what we need
to do for many.
[Sunday 12 Feb 2012] [22:50:14] <toad_> evanbd: i summed them up as basically 4
problems - 1) pitch black, 2) social feasibility of a darknet, 3) feasibility of
captchas, 4) fea
sibility of scalable fast-enough WoT-based chat, 5) relative ease of a regime
blocking freenet, no matter what we do re stego
[Sunday 12 Feb 2012] [22:50:54] <toad_> pitch black = technical feasibility of
darknet re security; there are also issues with scalability of darknet
[Sunday 12 Feb 2012] [22:51:32] * TheSeeker argues to change focus to
preserving what freedom exists, rather than trying to guarantee safety or
useful purpose for the a
lready criminally oppressed
[Sunday 12 Feb 2012] [22:51:36] <evanbd> 1) I don't think we've proven
we can get good topology on opennet, never mind deal with an adversary on darknet 2)
I think has to
start with better brand awareness (aka large opennet)
[Sunday 12 Feb 2012] [22:51:46] <toad_> #2 and #5 are the most obvious though -
if things are bad enough for enough people for a darknet, then they're bad enough for
the state t
o crush it, even if they have to put a dollar a month on the cost of an
internet connection to do it
[Sunday 12 Feb 2012] [22:51:55] <evanbd> 3) is hard, but other people
may solve for us
[Sunday 12 Feb 2012] [22:51:59] <evanbd> 4) seems doable
[Sunday 12 Feb 2012] [22:52:06] <toad_> imho re #1 we have good topology, there
are other problems; the graphs mostly look good iirc?
[Sunday 12 Feb 2012] [22:53:00] <toad_> evanbd: i'm skeptical that #4 is
feasible even for 10k users of such a system - FMS isn't that big
[Sunday 12 Feb 2012] [22:53:13] <toad_> WoT certainly isn't
[Sunday 12 Feb 2012] [22:54:46] <toad_> evanbd: five impossible things before
breakfast?
[Sunday 12 Feb 2012] [22:54:54] <evanbd> Yep :)
[Sunday 12 Feb 2012] [22:55:08] <toad_> if you've done five impossible things
before breakfast, why not round it off with a trip to the restaurant at the end of
the universe...
[Sunday 12 Feb 2012] [22:55:13] <evanbd> toad_: I think the long-term
problems are largely tractable or avoidable. But not easy.
[Sunday 12 Feb 2012] [22:55:27] <toad_> evanbd: i see no evidence that any of
those 5 are tractable
[Sunday 12 Feb 2012] [22:55:49] <toad_> evanbd: and the political/moral aspects
have become enormously more complicated in the last 2 years thanks to bitcoin
[Sunday 12 Feb 2012] [22:55:53] <evanbd> toad_: What they need is more
developers. And more math people. Which, imho, requires more brand awareness, and a
big enough user base to make smart, passionate people think it's worth their time.
[Sunday 12 Feb 2012] [22:56:01] <TheSeeker> toad_: build it and see, don't
give up until everyone runs out of ideas ;)
[Sunday 12 Feb 2012] [22:56:26] <evanbd> toad_: I don't have any
evidence either, but giving up seems like a bad approach :)
[Sunday 12 Feb 2012] [22:58:03] <evanbd> So anyway, what I was trying to
say is: usability and publicity are things we should be worrying about. And
attracting new users matters both in general, and *specifically to the security of
our existing users*
[Sunday 12 Feb 2012] [22:58:25] <evanbd> And therefore it's reasonable
to work on things like usability and publicity even though we have outstanding
security problems.
[Sunday 12 Feb 2012] [22:58:59] <toad_> evanbd: perhaps, perhaps not. cost of
comprehensive surveillance is linear at worst with network size, and a bigger network
may be a more tempting target.
[Sunday 12 Feb 2012] [22:59:13] <toad_> evanbd: especially if we were to
maintain a pretence that freenet is really secure
[Sunday 12 Feb 2012] [22:59:30] <toad_> evanbd: no ideas
[Sunday 12 Feb 2012] [22:59:32] <evanbd> "more secure than not using
freenet" seems like a good selling point :)
However, out of the below, which are a superset of the above, most seem
tractable, and one or two are either social problems which may yet turn out to
be tractable, or are problems the wider world also has to deal with. Whether we
have enough developers etc remains to be seen.
Security
--------
Fix Pitch Black: We're pretty sure Oskar's solution will work. Our friend
hasn't got back to us with simulations proving this, I may need to do them
eventually. Note that this isn't the only problem with swapping - the current
algorithm doesn't scale well, but Vive has a partial solution for that or at
least a research direction.
Fix Mobile Attacker Source Tracing: Several proposals dramatically reduce
vulnerability, including random routing and restrictions on routing to newish
nodes at high HTL.
Fix opennet: Opennet is a lot safer if MAST is difficult. It's always going to
be vulnerable to some degree, but maybe we can quantify that.
Blocking: If a country wants to block Freenet and doesn't care about collateral
damage, or is prepared to spend somewhat significant sums and bill the
consumer, we can't prevent them blocking it. See e.g. Iran, North Korea, Cuba.
However, it's looking like it'll be a long time before that happens in most of
the western world. If we can make some serious progress and get more users, it
may not happen at all, and Freenet becomes a long-term safeguard on the freedom
of the internet. In other words it's not hopeless.
Forums
------
How to make fast spam-proof distributed scalable forums: With the new key type, and some
sort of per-board announcement, I am convinced that we can do this, for "good
enough" scalability. The number of regular posters on a board is likely to be
limited in practice, and that helps.
Bootstrapping/announcement onto a forum: This one really is problematic - but
it's a problem that the rest of the Internet shares. The basic problem is
CAPTCHAs suck; they can often be cracked by computer, and even if they can't,
solutions can be bought by the thousand very cheaply. One specific problem
Freenet has is that we allow multiple attempts at the same CAPTCHA, because we
use them to create keys. I don't see any immediate solution to that.
Darknet
-------
We need to grow a darknet. This is a social problem as well as a technical one.
It may yet be intractable. But we know what we need to do technically, at least.
Performance
-----------
This one (raw speed, data persistence) may really be intractable, in that I
don't know how much more speed we can get, though I suspect we might get a lot
more data persistence. But it's not terrible, and we have a lot of ideas on how
to improve it, for example there are proposals for ways to probe how exactly
content falls out of the network, and our probe data suggests there is a
problem with link lengths, which we should be able to investigate.
Disk I/O and general system overheads: There are some important suggestions on
this, it should be possible to make massive progress, we probably can't
eliminate all periodic disk I/O but we can avoid writing many times every
second.
Bandwidth requirements: This is a serious problem, especially in places where
Freenet might be most useful. On the other hand, the situation may be improving
a bit.
Uptime requirements: See darknet, general performance; for darknet, FOAF
connections will help, for opennet, it's a matter of general performance. There
are various additional possibilities on darknet for long term requests.
Hardware Freenet Boxes would really help; router/home server boxes and cheap
hobbyist systems are rapidly approaching the point where they might solve the
uptime problem.
Fast distributed search
-----------------------
Right now even not-really-distributed search doesn't work very well, or isn't
official, or probably doesn't scale well. However, IMHO the problem is
tractable. There are various older proposals e.g. to limit the number of blocks
we need to fetch, while still preserving the indexes, and in the long term new
key types may allow for things that previously seemed impossible, such as
collaboratively maintained yet spam-proof indexes.
Transport plugins
-----------------
This was in the category of "technically feasible, great idea but will never happen
because there are always more important things to do". However chetan has made major
progress on this (for UDP-like transports for now), and hopefully it will be merged soon.
Greets to the estate of Douglas Adams. ;)
_______________________________________________
Devl mailing list
Devl@freenetproject.org
https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
