I assume you failed to read the what
I posted, for you seem to be unaware that (both I2P and Tor admit on their websites that) timing attacks only require the adversary to control the router at the entry and exit nodes in order to de-obfuscate the IP address of the user (or possibly for several users in the mix depending on the statistics of the low-latency and other factors), there is no need to invade the user's computer with exotic methods you mentioned.
Even hackers can (sometimes) do that.
I understand perfectly well what you've written, you seem to have misunderstood what I have written.
You have apparently not understood the implications that I was aware of since at least August 2013.
First, stop talking about Tor, we've never mentioned Tor and have already rejected it due to the problems inherent in their use of exit nodes.
If you don't want an exit node in Tor, then run a hidden service. Then the server at the onion layer hope preceding your hidden service doesn't know you are consuming the communication or are another onion layer hop that will forward on.
However a global adversary can still use timing analysis to determine that you are not forwarding communication and deduce you are running a hidden service.
Now: we are not aiming to prevent a global adversary who has control of all bordergates at all ISPs on the planet from making assertions and observations. We are aiming to prevent them from knowing whether Monero is running or not. They can know I2P is running, that we don't care about, but they should not be able to determine whether the person is harmlessly browsing the web or using Monero without completely decrypting the traffic.
This is why I mentioned that holistic design is so important. Because you've got several design variables that come into play on this.
1. The government can Sybil attack (inject nodes into) the network to gain decrypted communication and use this along with timing analysis given their global view of decrypted traffic across the nodes in I2P to backtrack which nodes in the I2P network are sending Monero communication.
2. As I explained in prior post, if the bloat of your ring signatures contributes to pool centralization, then if (as for Bitcoin) you end up with one or two pools with > 50% of the hashrate, then the government can national security gag order those and effectively decrypt and view the traffic. Since all mining nodes (e.g. pools) need to communicate with each other, the adversary can backtrack to which nodes in the I2P network are sending Monero communication.
3. The minimum block period (which is currently 10 minutes in Bitcoin, i.e. no micropayments to change the internet from the Google ad funded morass back to individual freedom) is depending on the orphan rate which is depending on the propagation delay of solved blocks. Since mining nodes (e.g. pools) must communicate this information, then the use of onion routing is unacceptable.
You can have the mining nodes send communication from a different IP address than which they receive communication (and send this information directly to each other instead of onion routing), but this still won't alleviate the Sybil attack of #1 nor the centralization of mining of #2, i.e. the users sending transactions can't know which pools are compromised so you've achieved nothing compared to each user employing a VPN.
The only way I can see to truly hide from a global adversary whether the user is participating in a crypto-currency network, is for mixing network to be immune from timing analysis.
This is why I've said since August 2013 that low-latency mix-nets are the incorrect paradigm for crypto-currency.
Why? Because a wannabe "global police state" (without resorting to the exotic methods I described) cannot possibly go after millions or billions of people around the world under the control of regimes and governments of all sorts merely for using a communications system.
Yeah but a low-latency mixing network subject to timing analysis can't help you.
I understand that you're a terribly clever and opinionated chap, but don't assume that we are ignorant or that we have made decisions for no good reason. I2P integration is for a specific purpose and is fully expected to be a stopgap if its shortcomings are not solved over time.
I measure my abilities on actual accomplishments and 25+ years of observing my designs compared to others. For example I used to work with this guy (and
Tom Hedges,
Lee Lorenzen creator of
Ventura Publisher,
Steve Guttman the former Product Manager for Adobe Photoshop) in mid 1990s on what became and is still shipping today as
Corel Painter:
http://relativisticobserver.blogspot.com/2012/02/fractal-design-formula.htmlhttp://relativisticobserver.blogspot.com/2012/02/miracle-of-paint-can.htmlhttp://relativisticobserver.blogspot.com/2011/12/2011-year-of-steve.htmlhttp://relativisticobserver.blogspot.com/2014/06/interesting-persons-part-1.htmlhttp://relativisticobserver.blogspot.com/2014/02/the-new-brand.htmlHere
I am interacting with Mark on this blog. And here I put
my entire life story on his blog. Oh my, I must have been very sick at that time.
Edit: I see
I was developing on Mark's blog what became "The Universe" essay on my blog linked in my signature. I see
I told him I am descended from the 1st governor of Kentucky and a ranking officer of the militia in the American Revolutionary War and the War of 1812 (also have warrior bloodline from Cherokee native tribe which doesn't originate from Isaac).