From: Robert Currey (Robert.Currey_at_AHand.co.uk)
Date: 13/06/03 20:00
On Thursday, June 12, 2003, at 08:47 PM, Mark Harrison wrote:
> Robert, if you're going to make the arguments stick on this for
> potential GVC members I think you'd need to bring with you some pretty
> solid underpinning of the following. These are off the top of my head
> and there are probably a lot more.
Thanks for the interest. In case anyone thinks I am undermining the use
of ADSL where it is applicable, I'm not. If you want a simple service
almost like flicking a switch go ADSL. But there has been mention of
wireless in the GVCC group, so if after consideration wireless is the
choice for provision in an area this is where my argument should be
considered.
I am convinced no commercial operation can make wireless provision
work, there is a persuasive argument in 'Building Wireless Community
Networks' Rob Flickenger ISBN 0-596-00204-1. There is a problem in that
you can't guarantee wireless provision - the best you can do is test,
send an engineer round and see if you get a signal; then it might work
until leaves come out in spring or a bird perches on an antenna. This
is not good enough in a customer - provider relationship you pay your
money and expect a service. I am effectively saying that the gvcc
member wanting a switch flicking service can't have it via wireless.
But there are successful wireless networks, these are community led
endeavours. Users need not be technically minded but they do need a
fair bit of enthusiasm. Indeed the gvcc expects this level enthusiasm
to provide support to the community ISP users, ADSL or otherwise; this
was in Mark's presentation at the TIC.
Confining myself to a potential gvcc user who is enthusiastic and in an
area unserved by ADSL, the following contains my reasons for an under
lying open network.
a) If you want to use the network the same way as an adsl gvcc user.
You can use the open network as a transport layer to the gvcc closed
and paid for network. In practice in XP and mac OSX you fill in a
wizard with username and password exactly like modem dial up wizard but
instead of phone number you put in an ip address. The data that
travels over wireless between your computer and the gvcc server is
encrypted and a damn site more secure than the phone line leaving the
house (i'm not talking wep but pptp over ipsec.)
b) By you using an open network you increase your neighbours chance of
connecting with full freedom whether to also connect to gvcc or do
something more.
c) You have freedom to do something more if you choose.
To me its b) that is crucial and the reason for my plea.
Assuming a) Mark's issues are pushed to the gvcc to deal with.
On the open network it will be admittedly a wild frontier and probably
worth the 20 quid to do a) above and have a sandboxed environment.
>
> Security - prevention of hacking, monitoring of traffic, security of
> email, validation of users, viruses...
a) above (pptp and ipsec) passes makes the security weak link the users
machine and the isp.
>
> Abuse - bandwidth hogs, illegal activities and content, impersonation,
> paedophilic grooming....
A job for the police not isps. On illegal activities, Hebden Bridge
might be a hotbed of direct action on an open network you can really
only hold the activist responsible; if in gvcc closed network the isp
condoned (through inaction) the action isp(& its executive) can be held
accountable.
>
> Services - email, web space, secure ordering, static IPs....
simple - not provided. These are things you need to pay an isp for. But
with an open network a rival isp to gvcc could spring in to fill a need
the gvcc has overlooked
>
> Integrity of content - child-safe spaces, validity of assertions and
> information, minimisation of spam,
if you don't use an isp you take control of such things yourself.
> copyright abuse....
becomes a personal liability/choice. There is no authority other than
the law, which is as it should be, isps should not be restricting users
for fear of their own liability
>
> Integrity of service - what happens if your local node is switched off
> for the evening, the week, or moves away? What would contention ratios
> be like on key geographic locations. Who's providing the free backhaul
> to the internet? Can all 1000 of us share yours?
Backhaul not provided - you don't get something for nothing. Friends
could choose to share their internet connection subject to contracts(bt
allows 20 users), but everyone is a free agent. With good density there
*should* not be a single point of failure, there always is in AP
arrangements.
>
> Stakeholder benefits - it's free but what else. (For example, GVC
> intends to offer business to business and business to consumer
> transactions of benefit to all parties - with the intention that
> cross-subsidy from commercial traffic will ultimately make the network
> free, or near free, anyway. Who would be responsible for making this
> bullet-proof on an open network?)
Not bullet-proof but robust. The point is that nobody is responsible
(apart from oneself) that's the goal, that is really presenting an
alternative product, and it doesn't prevent the closed option being
used over the same gear.
>
> Organisational image - likelihood of being a vehicle for grants,
> investment etc using your model. Breadth of appeal to potential users
> -
> how do you convince people you're not just skateboarders wanting to get
> pedestrians off the precincts?. Potential for serious bandwidth
> integration of community networks. Potential for underpinning major
> local infrastructure development....
Comparison = linux. A pretty fair public image, support, business
opportunities etc
I'll leave it there if I may.
Robert
This archive was generated by hypermail 2.1.7 : 13/06/03 20:01 BST