[LINK] Wireless Broadband for Regional Australia

[Richard]

On 20/12/13 1:27 PM, Paul Brooks wrote:
> On 20/12/2013 8:34 AM, Tom Worthington wrote:
>> On 18/12/13 11:40, Paul Brooks wrote:
>>
>>> ... FTTdp model in the Strategic Review ... distribution
>>> point) is a pit at the bottom of the driveway - or more likely,
>>> attached to the side of a nearby power pole ...
>> If most householders are accessing their broadband via WiFi and Mobile
>> Broadband, could you use it as the link from the distribution point (DP)
>> in the street into the household?  That way no extra equipment would be
>> needed in the house and a service could be provided to mobile users in
>> the street, as well as households.
> Sure you could add some form of femto-cell function for augmenting the mobile cellular
> broadband networks, to serve nearby homes and walking-by pedestrians - but I suspect
> you'll have real problems with phones in cars driving by due to the extremely rapid
> cell-switching that would occur every few seconds. Phones in cars wouldn't finish the
> handshaking with one cell before moving into the next.
>
>
> I distinguish very separately household networks (generally cabled and WiFi), and
> public networks (cellular mobile broadband) - and it goes a lot further than the link
> technology. Household WiFi is generally a private network, with no bandwidth/volume
> charges, relatively secure (on the household side of the firewall), and often relies
> on functions within the WiFi router to facilitate non-trivial apps such as NAT, port
> forwarding, VoIP proxies, multicast proxies, etc - which you lose in that model.
> A model such as you propose here (no extra equipment needed in the house) would:
> * be effectively forcing everyone (and every device) into the same security model as
> WiFi access at McDonalds, coffee shops, etc;
> * be useless for devices with cabled ports and not WiFi (think printers,
> set-top-boxes, DVD players, smart TVs etc),  and high-bandwidth devices such as NAS
> storage.
>
> So you'll still need a wired hub for these cabled devices, while forcing all access
> from one of your devices to the files on your NAS through a double-WiFi hop
> (remembering that WiFi is only half-duplex). If the data charging model was similar to
> mobile broadband, it would be unworkable.
>
> Most devices with just WiFi connectivity tend to assume there is a firewall/NAT device
> located on the other end of the WiFi hub - which would be missing in this instance -
> and you don't want to have to force your wifi-connected photo frame to have to jump
> through the web-based captive portal hoops that a tablet or laptop has to go through
> accessing coffee-shop/airport-lounge wifi systems.
>
> You lose the compartmentalisation that is important for home networks in limiting the
> scope of network broadcasts, particularly server advertisements. Imagine using the
> network browser to find a shared drive, and having to wade through all the services,
> servers, shared drives, network printers, DLNA sources and displays, etc etc located
> in all of the neighbourhood's homes! (and the security problems that might bring).
>
> The alternative might be to keep the WiFi-enabled broadband router in the home to keep
> the firewall and broadcast containment functions, with the uplink being also WiFi, or
> cellular mobile broadband to the pole outside - with all the performance limitations
> that brings.
>
> Personally, I think the initial assumption (most people are accessing their broadband
> via WiFi and Mobile Broadband) is an incorrect starting point. At home, people don't
> 'access broadband', they use broadband to 'access devices/servers/content' - its the
> same sloppy thinking that conflates "broadband" with "the Internet". In a home
> context, to an increasing degree much of those devices/servers/content is also located
> in their home and is not accessed over a public broadband link, and would have their
> utility killed if they were forced to be.
> Paul.
> _______________________________________________
> Link mailing list
> Link at mailman.anu.edu.au
> http://mailman.anu.edu.au/mailman/listinfo/link
>
I'll chime in on another point.

WiFi standards like 802.11n and the next one, 802.11ac, achieve their 
high speed by using MIMO - multiple in, multiple out - to provide 
spatial multiplexing. There are multiple antennas at the transmitter and 
the laptop (not, however, on mobile phones or tablets).

Taking a conservative calculation: if you have 4 x 4 antennas and each 
spatial path gets 50 Mbps - then the aggregate is 200 Mbps. OK so far ...

These standards just LOVE having lots of reflective paths around to 
choose from - such as you get inside a house.

They don't like a long distance between transmitters and receivers. At 
close-ish quarters, the four antennas may have a spread of 1 degree arc 
from the point of view of the receiver. At 20 metres, that's much less 
(if you want it worked out, find someone who remembers his high school 
trig!). The four antennas look like one - and with relatively clear air 
between tx and rx, you don't get the multiplication.

In other words: if you try to use modern WiFi in an environment that 
doesn't suit MIMO, you get the speed of a single spatial channel, to 
share between all users.

Richard Chirgwin