Re: [NNagain] A quick report from the WISPA conference

["Dick Roy" <dickroy@alum.mit.edu>]

[-- Attachment #1: Type: text/plain, Size: 1712 bytes --]

...

> 

> T-Mobile has signed up 1m+ people to their new Home Internet over 5G, and
all of them have really meaningful bufferbloat issues. I've been pointing
folks who reach out to this thread about cake-autorate and sqm-autorate, but
ideally it would be fixed at a network level, just not sure how to apply
pressure (I'm in contact with the T-Mobile Home Internet team, but I think
this is above their heads).

[RR] While there may indeed be a bufferbloat issue, it is also very possibly
a capacity problem.  T-mobile’s home internet offering is at 600MHz where
the maximum downlink speeds are around 17Mbps (aggregate) assuming that the
entire 35MHz of available spectrum is used and its about 10Mbps per user on
the uplink. It’s a little complicated because … . (This assumes QPSK
encoding and a coding rate around 0.7. You can double those capacity numbers
roughly for 16-QAM however the range will drop dramatically.) The point is
that when there are N users in a “sector” (beamforming can be used though I
am not sure whether T-mobile does or not), each user gets on average 17/N
Mbps downlink, and can compete for access to 10/N Mbps uplink.  If N is on
the order of 100, you can see that transmission rates are severely limited.
I have first hand experience of this.  When my download speeds come to a
screeching halt on my iPhome, e.g. when I am downloading an app from that
app store, all I have to do is look at the top of my screen, and EVERY TIME
it says … drum roll please … “Lucky you! You are now on our latest and
greatest 5G network!”  OK, it just indicates “5G”, but either way, I want to
throw this iPhone through some t_Mobile store window! :-):-)  

 

 


[-- Attachment #2: Type: text/html, Size: 3980 bytes --]