www.nettime.org
Nettime mailing list archives

<nettime> Don Cameron: The WiFi Dilemma
geert lovink on Sun, 29 Dec 2002 14:52:32 +0100 (CET)


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

<nettime> Don Cameron: The WiFi Dilemma


(fwd. from the community informatics list. perhaps interesting for armin
medosch, youarehere and other wireless network initiatives. ciao, geert)

From: "Don Cameron" <dcameron {AT} mudgeeab.com.au>
To: <communityinformatics {AT} vancouvercommunity.net>
Sent: Monday, December 16, 2002 6:31 PM
Subject: [CI] The Wifi dilemma


When the best of intentions destroys community growth and development
opportunities.

By all the indicators of cost efficiency, technical superiority, ease of
deployment and an ability to increase community ICT accessibilities;
Wireless (Wifi) networks 'should' be taking-over as the primary means of
deploying Internet services to the millions of needy and disadvantaged
peoples of our world... Yet other than a few showcase models offering
limited services and fighting for sustainability, most global Wifi
initiatives continue to slide into the ether region of ICT obsolescence -
joining all the other fantastic ICT projects that for some reason or
another, just never quite made it.

In a world where we adopt so much technology at a pace that sees start-ups
grow to global corporations overnight, where two or three mobile phones per
household is not enough, why then is the Wifi (with so much obvious
potential), languishing, and in many cases slipping further into a spiral of
decline?

Supporters of Wifi's are quick to blame Governments and Telco's for a lack
of support, yet rarely do they look inward to see how their own models and
modes of operation stack-up against the realities of supply and demand. What
does a Wifi really offer? Does it meet the requirements of the market? Are
the services sustainable within the available supply-chain model? How does
delivering last-mile services via Wifi impact on upstream service providers?

No ICT business will survive for long without addressing these fundamental
questions.

A Wifi can provide any computing service, but what is most desired by
markets is access to the Internet. It is doubtful that any Wifi will survive
long-term without offering this as a core service.

Most Wifi administrators acknowledge this need, yet few seem interested in
constructing a business model supportive of equitable 'net deployment. Most
want to provide their customers with access to the 'net for free and scoff
at the concept of entering proper fee-for-use arrangements. "A domestic
household can buy a 512K connection for $50.00 a month, so why can't we buy
this same connection and distribute it amongst 10 households for the same
price, it's perfectly feasible" is the argument most often voiced by
frustrated Wifi administrators.

Honourable and altruistic?... certainly... Correct?... to anyone with an
understanding of the mechanics of bandwidth deployment, "Yes"... to anyone
with an understanding of the economics of bandwidth deployment... sadly,
"No".

The worst aspect of this altruistic drive to provide free 'net access is
that it effectively precludes the Wifi from being a viable means of
deployment and subsequently deprives millions of people from access to this
vital ICT service.

The mechanics of bandwidth deployment is relatively simple providing you are
not a mathematician (because the sums simply don't add up). Telco's offering
broadband do not have inbound pipes capable of supporting every outbound
pipe. A Telco with 10,000 customers on 512 K broadband will not support this
with a 5,120,000 K inbound pipe - the costs would be astronomical - what the
Telco does is to assess bandwidth usage trends (usually 5-8% domestic and
12-15% commercial) and provides sufficient inward bandwidth to meet the
projected demand. Telco's know that very few customers will come close to
100% bandwidth utilisation, and that a great many more will fall below 1%
utilisation. Hence the 'rule' is to purchase inward bandwidth at a maximum
level of around 15% of all possible outbound connections.

None of this is really any different to the town planner tasked with
determining an appropriate size for an urban water reservoir. The assumption
is that most people do not run their garden hoses 24x7 for 365 days a year.

This formulae for equitable bandwidth distribution prevails right across the
Internet and is common amongst most ISP's and other network access providers
seeking to provide access at a realistic cost.

Once the mechanics are understood, the economics of bandwidth distribution
also becomes clear - costs are structured on the basis of a maximum of about
15% bandwidth utilisation. Some customers use more than this (good for
them); a great many others will use less (largely unaware they are
subsidising the heavier users). Costs are structured so as to return a
profit on the bandwidth required to sustain the entire customer-base.

So what happens when we bring an altruistic Wifi into this model of supply
and demand?

Wifi administrators know only too well that it takes perhaps half-a-dozen
512K broadband connections to serve a customer-base of 40-50 broadband
subscribers, most of whom will still be able to connect at 512K (adequately
supported by Cache Servers and other means of minimising bandwidth
utilisation) - this is because the Wifi can similarly work on the 15%
deployment "rule". Very few of their customers will ever achieve (or want)
100% bandwidth utilisation. Yet these same administrators expect upstream
providers (Telco's etc.) to provide them with 100% utilisation of inbound
pipes at the same cost as residential customers perhaps utilising less than
5-6%.

Obviously the threat to Telco's from Wifi's is not that they compete against
core services; it is that Wifi's seek to utilise 100% of inbound bandwidth
(subsequently distributed to a larger customer-base) thereby requiring a
completely different costing structure to sustain.

A Wifi can be a very successful, and perhaps even an essential method of
deploying ICT's to rural and remote communities when Wifi administrators
acknowledge the need for upstream providers to similarly remain viable. This
is achieved by providing upstream providers with income commensurate with
the services deployed. Wifi customers should pay the same as any other
broadband recipient; and Wifi administrators should pay for the bandwidth
they use. We are all part of a connected loop in this global village and
Wifi's will never thrive and properly support our communities until they
cease to view the world in isolation. Bandwidth costs money. There is no
shortcut to negate this reality.

Best regards, Don

--

From: "Tom Abeles" <tabeles {AT} attglobal.net>
To: <communityinformatics {AT} vancouvercommunity.net>
Sent: Sunday, December 22, 2002 8:33 PM
Subject: Re: [CI] The Wifi dilema

Hi Don

let's look to the future. What happens if we abandon all that fiber and make
the
entire system wireless and meshed? Universities spent a lot of money to wire
the
dormitories and the entire campus. Now, with cellular, some universities are
thinking about abandoning their twisted pairs in the dorms for lack of use.
And
with a campus being wireless, the campus network becomes obsolete. With
meshed
systems, large communities would seem to have service until they come up to
the
fiber.

Right now, there is consideration about "unlicensing" all that bandwidth
which
the US government auctioned off. Digital seems to expand the capacity
without
interference, and apparently giving almost infinite room across the spectrum

thoughts

tom

--


From: "Don Cameron" <donc {AT} mudgeeab.com.au>
To: <communityinformatics {AT} vancouvercommunity.net>
Sent: Monday, December 23, 2002 7:04 AM
Subject: Re: [CI] The Wifi dilema

Hi Tom,

I suspect most university networks have more in common with a corporate LAN
than they do a community Wifi seeking to distribute free public access to
the 'net - Universities (like businesses) usually provide a 'closed loop'
network service to a known and identifiable customer-base; who in turn pay
for services through tuition fees. Universities rarely open their networks
to the world at large, nor do they attempt to act as an ICT supplier to
disadvantaged members of our communities.

My employer has a not dissimilar networking requirement to many
Universities - we are a business physically spread over 200 hectares with a
dozen on-site buildings & facilities. Staff regularly move from building to
building necessitating a need for my section (IT Services) to provide common
points of access throughout the site. Wireless networking provides a very
good means of maintaining network connectivity to mobile staff reliant on
the use of Laptops, Data Loggers and inventory management scanners etc.

Your post leads to a few interesting questions over the technical
possibilities - as well as (more interestingly) the use, or lack of, of your
current network... (re your comment "some universities are
thinking about abandoning their twisted pairs in the dorms for lack of
use").

At the technical level, assuming your current twisted pair network is Cat 5E
(and that unlike us, you are not prone to violent electrical storms!); have
you considered upgrading your network switches to Gigabyte Ethernet
supported by VLAN and server trunking? (multiple Cat5 connections to your
major network servers). This offers comparable performance to fibre at about
one-tenth the cost because you are not required to re-wire the entire site.
Gigabyte transfers are quite possible over properly installed Cat 5E. This
can be easily supported by a low-cost wireless LAN for mobile users possibly
meeting your current and projected networking requirements.

The cited 'lack of use' of the physical network really leads to questions
over the viability of any proposed network expansion (fibre or Wifi). Why
expand if your users are preferring to use WAP over in-house network
services? A few months ago I wrote on this trend in Australia - how a great
many of our youth were abandoning the 'net in preference to WAP and
text-based (Cellular) communications. The trend seemed to be (and continues
to be) that Cel-comms is increasingly preferred for 'groupware' style
community communications supported by the 'net for file-sharing and research
applications.

With the increasing ease by which some on-line services are available over
Cell (Email and basic surfing capabilities); perhaps your users are
suggesting that expanding your in-house network is really not seen to be a
need - that the future of ICT communications lies in a different direction.
Has any analysis been undertaken on the implications of this for community
ICT installs?

Best, Don


#  distributed via <nettime>: no commercial use without permission
#  <nettime> is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: majordomo {AT} bbs.thing.net and "info nettime-l" in the msg body
#  archive: http://www.nettime.org contact: nettime {AT} bbs.thing.net