Drak [ARCHIVE] on Nostr: 📅 Original date posted:2013-12-08 📝 Original message:On 8 December 2013 12:37, ...
📅 Original date posted:2013-12-08
📝 Original message:On 8 December 2013 12:37, Luke-Jr <luke at dashjr.org> wrote:
> Encryption is useless here. We want everyone to be able to download Bitcoin
> clients. Binaries on sourceforge are signed by multiple parties using
> gitian.
>
> > Decentralization:
> > So long as we actually use DNS, the website is centralized :( However,
> > its content isn't (can be forked on GitHub), but regarding the domain
> > name, there is not much we can do against this AFAIK.
>
> So long as someone has root (or a user that can modify it), the website is
> centralised. To really solve this, we would need a dedicated server that
> accepts commands only when signed by N-of-M parties, inside a cage locked
> by
> padlocks with keys held by independent parties, with a SSL certificate
> issued
> by an authority that has multiple parties watch it every step of the way
> into
> that server.
Malicious actors with root access to the server is another issue entirely.
Sure it's a problem, but it is not an argument not to have a properly
signed SSL certificate.
With out one, the exploit can be performed on routers to redirect traffic
through a third party alter the content of the site (like the links on
bitcoin.org to various wallet projects) and then onto the correct
destination. SSL at least mitigates that. For example it would be trivial
to impersonate Electrum's site or whatever, "change" the link on the fly
that appears on the trusted source bitcoin.org via BGP redirection. Now
users will be directed to the scammers site which could be identical except
for domain name and of course malicious binaries.
BGP redirection is a reality and can be exploited without much
expense/effort. MITM is a real world threat, not some theoretical
possibility - reports show it's happening on an unprecedented scale. SSL is
essential - that's a no-brainer. Sure other measures are important, but
without SSL there is almost no point to any of the other options.
SSL is so considered so important that the *HTTP 2.0 spec might be SSL
only*according to recent discussions at the W3C (
http://lists.w3.org/Archives/Public/ietf-http-wg/2013OctDec/0625.html).
Drak
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20131208/c6778642/attachment.html>;
📝 Original message:On 8 December 2013 12:37, Luke-Jr <luke at dashjr.org> wrote:
> Encryption is useless here. We want everyone to be able to download Bitcoin
> clients. Binaries on sourceforge are signed by multiple parties using
> gitian.
>
> > Decentralization:
> > So long as we actually use DNS, the website is centralized :( However,
> > its content isn't (can be forked on GitHub), but regarding the domain
> > name, there is not much we can do against this AFAIK.
>
> So long as someone has root (or a user that can modify it), the website is
> centralised. To really solve this, we would need a dedicated server that
> accepts commands only when signed by N-of-M parties, inside a cage locked
> by
> padlocks with keys held by independent parties, with a SSL certificate
> issued
> by an authority that has multiple parties watch it every step of the way
> into
> that server.
Malicious actors with root access to the server is another issue entirely.
Sure it's a problem, but it is not an argument not to have a properly
signed SSL certificate.
With out one, the exploit can be performed on routers to redirect traffic
through a third party alter the content of the site (like the links on
bitcoin.org to various wallet projects) and then onto the correct
destination. SSL at least mitigates that. For example it would be trivial
to impersonate Electrum's site or whatever, "change" the link on the fly
that appears on the trusted source bitcoin.org via BGP redirection. Now
users will be directed to the scammers site which could be identical except
for domain name and of course malicious binaries.
BGP redirection is a reality and can be exploited without much
expense/effort. MITM is a real world threat, not some theoretical
possibility - reports show it's happening on an unprecedented scale. SSL is
essential - that's a no-brainer. Sure other measures are important, but
without SSL there is almost no point to any of the other options.
SSL is so considered so important that the *HTTP 2.0 spec might be SSL
only*according to recent discussions at the W3C (
http://lists.w3.org/Archives/Public/ietf-http-wg/2013OctDec/0625.html).
Drak
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20131208/c6778642/attachment.html>;