Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
YeomanNet
24 views
Skip to first unread message
Chris Riley
Jul 3, 2014, 12:41:49 PM7/3/14
to dev-decent...@lists.mozilla.org
Hi all,
At Tarek's suggestion, I'm sharing a year-old white paper of mine with
this list, in case it is of interest - attached and available from SSRN
<http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2462214>.
It's just a concept piece, something I first started thinking about ~6
years ago when Twitter came online and had a ton of fail whales. The
recovering computer scientist in me said, "why on earth is anyone still
using centralized storage at scale" (though of course it's not truly
centralized), and started thinking that my dissertation research might
apply... I was studying the theoretical benefits of adding an assumed
reliable, but constant-cost supervisor to a DHT, and finding some
provable though not earth shattering benefits (basically because the
supervisor can maintain a balanced tree overlay relatively easily). I
never did anything with that, but when Prism came around, it surfaced in
my consciousness again, as a way to build a system resilient to hidden
mass surveillance that still allows for (sort of notice-and-takedown
style) moderation of things like IP theft and porn.
This is all a looooong ways away from reality, but I think it's still an
interesting idea, and am curious what others think.
Chris
At Tarek's suggestion, I'm sharing a year-old white paper of mine with
this list, in case it is of interest - attached and available from SSRN
<http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2462214>.
It's just a concept piece, something I first started thinking about ~6
years ago when Twitter came online and had a ton of fail whales. The
recovering computer scientist in me said, "why on earth is anyone still
using centralized storage at scale" (though of course it's not truly
centralized), and started thinking that my dissertation research might
apply... I was studying the theoretical benefits of adding an assumed
reliable, but constant-cost supervisor to a DHT, and finding some
provable though not earth shattering benefits (basically because the
supervisor can maintain a balanced tree overlay relatively easily). I
never did anything with that, but when Prism came around, it surfaced in
my consciousness again, as a way to build a system resilient to hidden
mass surveillance that still allows for (sort of notice-and-takedown
style) moderation of things like IP theft and porn.
This is all a looooong ways away from reality, but I think it's still an
interesting idea, and am curious what others think.
Chris
0 new messages