On 9 December 2014 at 14:10, Christian de Larrinaga <[email protected]>
wrote:

> Hi Gord,
>
> ? http://www.artemis.com/pcell
> Be interested to learn if anybody has got under the veil of this one?
>

I haven't seen that before, but I can guess having skipped through the the
"tech" vid in a minute or less. What follows is a common rant, tailored to
the  "wow technology!" being pushed|sold|wavedabout|linked-to ....


Instead of quite a lot of radio station sites oriented in cells, which
works OK until capacity is reached in that service area, they will use lots
of little radio stations, effectively forming smaller and smaller cells.
Nowt new - some cells are tiny, like femtocells. Some users use P2P methods
if they want the same things that other do. Nowt new BitTorrent or VHS copy
sharing are examples. To share effectively means proximity, awareness and
similar needs.

Every single radio station of any form that had capacity or throughput
issues has used a similar technique since day one. Not all had good
marketing.

Reduce the range and you improve co-channel interference (as well as the
side effect of reduced adjacent channel interference due to lower EIRPs,
both wanted and spurious). As long as each new small, local site has not
hit maximum capacity, for the required users in that service area,
everything is sweet.
A soon as a limit is reached, you have to reduce the service areas, with
reduced powers (which doesn't always mean reduced performance as you might
think), by adding more cells. The process repeats until someone has a good
idea. Limit hit, add more sites or upgrade tech to faster or wider or
$something_normal.

Every now and again, someone has a good idea. Some of these need marketing,
some gain traction own their own merits, some are obvious as soon as you
see it. Some are purely theoretical and may take decades and several
generations of kit to be practical.

In the days of morse and trained telegraphists, a station could hit it's
capacity in various ways. Staffing limits (morse training takes time and
good fists take decades to become good), lack of telegraphy equipment and
links, power and cooling constraints (watercooled sub-megawatt stations
existed in many countries), paperwork/procedural inefficiencies, weather
and climatic problems. Originally, none of these where limitations of the
"ether" as the transmission medium was called. Eventually, there were so
many stations that more and more frequencies, or groups of them, known as
"bands", needed to be used, as well as some form of planning to make sure
that the could co-exist without saturating the medium. Lower power, shorter
range stations were introduced, in a similar method to nano-cells. They
were shorter range not really by design, but by chance, unused bands that
were less suitable to long range working could be utilised for shorter
range use at certain times of the day.

At this point in history, we have frequency division (different frequencies
and bands), time division (operators allowing other traffic on a
time-sharing basis), range-limiting (where long range working wasn't
needed), backhaul limitations (onward transmission by telegraph cables),
procedural optimisation (streamlining message handling), beamforming (using
one or more directional antennae and exploiting them actively, often by
human intervention).
At this point in the tale, we're looking at the late 1920's and early
1930s.

In the remaining 90 years, throughput speeds have increased, equipment has
become more portable, cheaper in relative terms for the end-user
(arguable), all-pervasive and an expected part of everyday life.

The problem of the 20's and 30's are the same today.
User expectations, which should be high, are in effect destroyed by bad
experience of mobile communications especially in busy areas. Cells are
large due to lack of investment and forward thinking. Backhaul is a
problem, as it is everywhere, until you have enough. If "enough" backhaul
can't scale up when demand goes up, you hit the problems when the limit is
reached.

The actual radio media (the "ether") hasn't changed, the same laws of
physics are the same as they were 100 years ago, and probably have always
been similar to what we see today (or think we see), we (RF people as a
subset of tech society) may understand them better, that's all. The obvious
solution to most of the perceived problems are to use diverse backhaul
(install femtocells on xDSL all over the place), or massively diverse
backhaul (use technologies like bittorrent or similar P2P) to provide more
and more sites (cells) to provide the service to the users.

Nearly 90 years ago the same thing happened:-

More radio stations were installed where needed to provide capacity as
needed, not just fill in gaps in coverage
..and..
Radio ops would share the weather and traffic lists and even rebroadcast it
to stations further away or with less advanced equipment, using diverse
methods.


That makes 3 assumptions - customers had to be willing to pay, companies
wouldn't overspend or under-provide and that weather was wanted by many
people.

he selling point of this pCell $product|concept seems to be based on the
assumption that everyone wants the same Netflix movie at roughly the same
time. It claims to be able to provide 100% capacity to every user. As long
as they want the same Netflix movie.

I'm not convinced that the speaker knows the physics or understands the No1
problem he's creating for the customer - if user 1 wants to watch porn and
user 2 doesn't, then user 3 had better have the porn that user 2 wants or
the 100% claim is bullshit. You can only have 100% once, not twice, if the
content is different, in a given spectrum and coverage area. Yes, the
coverage areas are small (probably near-field). Yes, it's new to him. It's
possibly new to the audience. There are many parallels to what he|they have
claimed to invent|find|do.
Hell, the POTS telephone system itself is a possible analogy back to the
days of dial-a-disc and speaking clock. Limited content in the case of
dial-a-disc or speaking clock. Shared information at the end of the link
(you set your watch by it and could tell people that asked you the time
thereafter). Backhaul and coverage was extended until almost every building
in the country was served by a local dissemination station that provided
the service. Humans provided the P2P bit at the end. Near-field, local
comms, were by acoustic dissemination between people who wanted the
information/message/content. Service area is limited by the laws of physics
(transmission lines, EMC issues, power). You can optimise the exchanges to
improve speed (STD).

For datacomms over RF we have come so far as a society from those early
days that the users are far, far removed from the technology that they
can't understand the physics or identify the limiting factors in their
troubles.
It's an exciting technology in the same way that blue smarties were
exciting. They are still smarties and are fucking chocolate inside. There
are only so many in a box. If you want, you could cut them up and move your
mouth closer to eat them. You could pop one in your mouth and share with
your neighbour if they want one. Smarties can't be duplicated. You cannot
break the laws of physics, or smarties.

Forgive me if I chuckle at the audience at the end. Either it's a polite
audience or he's completely baffled them. I give 10/10 to the heckler, the
"tech" not so much.

The team are obviously having fun and I don't want to piss on that for one
minute, but it's hardly a breakthrough in historical terms as well as
communications technology terms. I wish them well. I feel sorry for people
who buy into that expecting to use it to solve a problem unless it's a
netflix problem AND everyone wants the same.
Enthusiasm for something like that is hard to exude when the only TV-style
media I watch are MST3K shows from the nineties. I doubt anyone nearby me
would benefit even if I did rebro them locally at low power. I accept that
groups of people tend to want to watch the same things, sometimes at
roughly the same times. They tend to group closely together. That's a
software opportunity for delivery efficiency, certainly not an RF
breakthrough.

There is only one 100%, even using "means diverse" as we all it. You cannot
provide 101%, nor can nature, without some convenient assumptions. I wish
he'd found an unlimited number of universes we could parallel up by using
bonding, *that* would help. Actually, I wish he'd found a better
compression algorithm for Netflix movies.


PS: Sorry, due to ranting nature of this I haven't spell-checked it or read
for continuity. Several phone calls interrupted it so read between the
lines and paras to get the sentiment.

-- 
sent via Gmail web interface, so please excuse my gross neglect of
Netiquette

Reply via email to