On February 23, 2014 2:11:26 AM EST, Collin Anderson <col...@averysmallbird.com> wrote: >On Sat, Feb 22, 2014 at 9:35 PM, Nathan of Guardian < >nat...@guardianproject.info> wrote: > >> I just hope that the "Tor is slow" meme, along with >> other famous ones like "PGP is hard" and "No one cares about >privacy", >> are continued to be challenged, tested and not accepted as >inevitable. >> > > >Certainly, however, I am testing this meme now, and it has validity. I, >right now, pulled the latest version of Tor to a Linux machine in Iran >and >did a comparative measurement of throughput with and without. Although >the >download started out reasonably fast, after a few seconds it seems >probably >that my connection was throttled down. Whereas without Tor, my >throughput >averaged 84.81 Kbps, with it is currently 14.66 Kbps and rapidly >dropping. >If anyone doubts that Tor's overhead is that high, at some point I >could >keep tearing down connections and narrow the likelihood of a poorly >performing bridge. However, factually, Tor is slow.
I just ran the same test with Psiphon3 and my throughout was 0.0kbps... (I joke!). I do appreciate your effort to add hard data to the thread. My point was not that Tor is not slow (it is, and generally 1/4 the bandwidth as your tests have shown), it is that the meme overrides any reality of possible usefulness, even when anonymity is not the user's goal. I was simply arguing for Orbot to be on the list of possible solutions that were being promoted, because especially on mobile, use of async apps that do their networking in the background mitigate the speed issues greatly. Here is my own direct data to share. On 9/11 in lower Manhattan, nearly all communications systems failed due to being taken out in the WTC collapse or overloaded by the mass amount of people trying to find each other. This included terrestrial internet links, again due to local routing through switching sites downtown and loss of power, etc. The only network that remained viable was Mobitex, a mobile pager data network that had been written off as too slow for the mobile web. Fortunately, RIM realized it was the perfect network for their early Blackberry devices, so long as they built their messaging system and other apps to mask the time it took for a message to send or be received. Web browsing on the early Blackberries was a joke, even with stripped down WAP sites. However, having a Blackberry (and a Palm VII also mobitex based) below 14th street on 9/11 meant that I was able to stay in touch with the small percentage of friends (.3% of new Yorkers perhaps?) who also had mobitex-based devices, I could still send and receive email, and that I became a source of information and coordination, while everyone else was disconnected and offline. (I also deployed a working WiFi mesh later that week, but that is for another thread!) If Tor is harder to block (which it is), more sustainable to operate (which it is), better for protecting users privacy from the operators (which it is), then my hope is that we see an increase in use among apps that are well designed for the type of network it is. That will only happen if those of use who see the value of Tor (or any circumvention service that prioritizes privacy above speed) make sure the apps are using it wisely. In this golden age of mobile messaging, may all app designers channel their own inner Blackberry for the benefit of all users! +n -- Liberationtech is public & archives are searchable on Google. Violations of list guidelines will get you moderated: https://mailman.stanford.edu/mailman/listinfo/liberationtech. Unsubscribe, change to digest, or change password by emailing moderator at compa...@stanford.edu.
