[tor-talk] Why so may exit node IP addresses in the TBB?

Juan Miguel Navarro Martínez juanmi.3000 at gmail.com
Fri Jul 22 16:20:07 UTC 2016


It answered their question about "When it started". And I think the
explanation should be on the 4.5 release page[1]:

"""
Usability Improvements

[...] Tor Browser is also much better at handling Tor Circuits in
general: while a site remains in active use, all associated requests
will continue to be performed over the same Tor Circuit. This means that
sites should no longer suddenly change languages, behaviors, or log you
out while you are using them.

[...]

Privacy Improvements

On the privacy front, the 4.5 series improves on our pre-existing first
party isolation implementation to prevent third party tracking. First
party isolation provides the property that third party advertisements,
like buttons, and "mashup" content that is included on one site will
only know about your activity on that site, and will not be able to
match it to your activity while you are on any other site. In other
words, with first party isolation, Facebook, Twitter, and Google+ can't
track you around the entire web using their infamous like buttons.
"""

[1]: https://blog.torproject.org/blog/tor-browser-45-released


On 2016-07-22 at 18:06, Allen wrote:
>>>> It used to be that, if you opened five sites in TBB, all five would use
>>>> the same exit node. Now it appears that each of the five sites has
>>>> the IP address of a different exit node.
> 
> It started on the alpha development of 4.5[1]:
>>
>> "This release features [...] isolation for circuit use. All content
>> elements for a website will use a single circuit, and different websites
>> should use different circuits, even when viewed at the same time."
>>
>> [1]: https://blog.torproject.org/blog/tor-browser-45-alpha-1-released
> 
> 
> That doesn't answer the question though why the different circuits use
> different exit nodes.  They could all use the same exit node.  I recall
> seeing fair bit of debate over how many exit nodes a Tor client should use
> and how often it should switch exit nodes...
> 

-- 
Juan Miguel Navarro Martínez

GPG Keyfingerprint:
5A91 90D4 CF27 9D52 D62A
BC58 88E2 947F 9BC6 B3CF


More information about the tor-talk mailing list