[tor-dev] Proposal 287: Reduce circuit lifetime without overloading the network.

[teor]

On 2 Dec 2017, at 04:56, Fernando Fernández Mancera <ffernandezmancera at gmail.com> wrote:

>> If everybody chooses different amounts of data, that makes them easier to
>> identify.
> 
> This reason is why it is not recommended to change the default value, same as for the MaxCircuitDirtiness parameter, so I will specify it on the proposal.

If a Tor option has a default value, and we don't recommend changing it,
we try to remove that option. So we are unlikely to accept a patch for a
new option unless there is some compelling use case for a group of users
to set it differently.

Otherwise, we would accept code to set a default and keep that default.

Onion services may be a use case for this option, because if too much data
gets sent through a single circuit, it makes it easy to identify the service's
guard. (Of course, this only protects the service if the next circuit uses a
different guard, which is not guaranteed.)

Alternately, Tor Browser may be a use case for this option, if it wants to
manage its circuits itself, regardless of the amount of data they send.

T