Re: [Linphone-users] Why Android (Oreo) phones, are actually less reliable with TCP vs. UDP

[Greg Troxel]

Juha Heinanen <address@hidden> writes:

> I charged my Android 7.1 phone in the evening and let Wifi on and
> baresip running.  In the morning about 12 hours later without touching
> the phone, I made call to its baresip account. Phone was ringing and I
> was able to answer the call.
>
> Then I checked what had happened to battery.  Baresip had used 3%, Phone
> Idle 3%, Android system 2%, ..., and WiFi 1%.  Battery also reported
> that at this rate there is 3 days left.

Thanks for the data point.  That does not surprise me, since I leave k-9
running.  That's IMAP not SIP, but it's basically the same thing as far
as network/wakeups in the idle case.

I wonder what would happen if you doubled or halved the registration
refresh interval.   But obviously you are in a pretty good place already.

> Based on this, push notifications don't make any sense whatsoever. They
> just add complexity and delay and are a security and privacy risk.

Agreed that they add complexity and delay.  With a non-proprietary
self-hosted push server, and N applications, you could perhaps have 1x
of the 3%, and not N * 3%.

I suspect that part of the drive to extreme power saving measures (by
extreme, I mean things that break functionality) is due to a prevalance
of badly behaving proprietary apps and a desire to protect the battery
from them.  With apps that only do what is in the user's interest
(rather than e.g. using energy for tracking and ads), and that are well
written, this is far less necessary.