Inkwell: Authors and Artists

    
Ha! I was trying to follow the book because I figured that's what you'd 
rather talk about, but if you want to get into questions of economic, 
political, social impact, we can certainly go for it.

Who pays for the infrastructure? One question is how much the 
infrastructure costs. Supposedly most of the costs for POTS and cellular 
both are in tracking and billing - the service itself is cheap. Probably 
true for WiFi, too - the commercialization overhead has the greatest cost.

It's hard not to think of WiFi as free or nearly free when that's how it's 
offered in so many cases.

But there are definitely costs. Who pays? Schlotzky's provides free 
wireless, and they say that it's cheaper to provide the service (as an 
adjunct to other uses of their T1 connections) than their cost for 
bathroom tile.

Are they making a lot of money with it? Hard to say, because part of the 
return for Schlotzky's has been in the press they've got for providing the 
service. But there's enough of an advantage that they're still rolling 
wireless access out to addtional stores.
  

    
(airman), it sounds like you're talking about operators of private networks
opeing their service out to other users, rather than businesses who install
hotspots to attract customers. The incremental cost of adding a hotspot is
minimal, but it's not free. And as I said earlier in this topic, the added
cumulative burden on the ISPs and backbone providers is something that they
may not have built into their traffic projections.

The business model for this stuff is still foggy, as far as I can tell.
  

    
From the WISP side, the Cometa model seems interesting: wholesale wireless 
to companies like McDonald's who can offer for free as an amenity (or 
potentially resell?). McDonald's was offering coupons for X minutes of 
connectivity with a purchase, which feeds a secondary business model, for 
tracking and metering.
  

    
The McDonalds hotspots make no sense to me at all. I'd always understood the
whole design of a fast food place was "eat it and beat it"--sightly
uncomfortable seating and colors, everything possible to turn over the
tables as quickly as possible. Offering WiFi seems exactly counter to that
kind of plan.

Ultimately, the wiFi world will have to be like the cellular telephone
network: you don't have to think about whether the local hotspot supports
your account. You just assume that it does, one way or another. Working
backward from that ideal to a system where the service provider makes a
profit is still a big gap.
  

    
I think the idea is to make McDonalds a more attractive place to make
a pitstop when you're on the road.  It's for checking your email, not
hanging about all day.
  

    
That's it exactly.
  

    
But it seems like it'd be awfully attractive to the 11-year old whose
parents won't buy him a DSL line. I'd expect the place to fill up with kids
downloading music files.
  

    
Shhh!
  

    
It's gotta be the most trouble-free way to provide Internet access to
the public.  All the hardware's in a back room where nobody but the
repair tech is allowed to touch it.  Way cheaper than a computer lab.
  

    

> as anybody seen a change in the way they relate to networking or the Inter-
> net because they've been able to cut the cable? Does it really make any
> difference to be able to go online from a Starbucks?


It has changed my morning routine and made my day a little more pleasant.
Instead of having coffee and reading the newspaper together and then heading
off to our separate offices for our morning online routines, my wife and I
now get online together.  That matters.

I was hoping to find out how much easier it is to get online while traveling
now that I have an Airport card, but my freakin' iBook just went belly up on
me for the second time in a month.
  

    
One of the complaints I often hear about WiFi is "Why don't they figure 
out how to get cellphones to work, first?" I've found myself in several 
situations where I couldn't make a connection, and I know that happens to 
everybody. The idea of mobile connectedness is just so compelling that 
people seem to take the hassles in stride.

I heard about a meeting of attorneys here in Austin. They all had their 
Blackberries but they were at a country club outside cell range, so they 
couldn't get a signal. One guy mentioned that he connected in the middle 
of the parking lot, so on break they all ran out the door and there were a 
dozen or so attorneys standing there, Blackberries in the air, catching 
email.
  

    
That's just an amazing image, Jon!
  

    
I use WiFi for a home network, both for an upstairs computer (which
let me get rid of Cat5 ables running out of windows) and for a laptop.
I find that I am not wireless with the laptop, though. Because of the
power draw of the wireless card, I get only and hour or so of battery
life, so I am pretty much tied to the power cable. I hope to go to a
Centrino box soon to remove this constraint. So it is nice to have the
laptop sitting in front of the couch so I cna check IMDB or the Well
while watching TV (I am an obsessive multitasker). 

My wireless net went in and remains pretty painless. I tried to
install one on my inlaw's network so they could share their cable modem
connection, but the base unit was running Win 98 with a ton of
problems and extraneous programs. After 2 days of work I never got it
running. Batting 500!
  

    
I'm jumping into the conversation late, I have been on vacation
and off line.  I have two laptops (one XP, one win2k) and an older
PC card.  Don't have it handy, but I know it was older because we
had to grab new drivers off the net because it didn't support XP.
So now I am looking at routers and access points.  One thing not
addressed in your book that is an issue now is 802.11g.  That would
seem to be the way to go these days.  The other thing that confuses
me when looking at catalogs is when some access points are advertised
at 22Mbps.  I thought b was 11Mbps and g was 56Mbps.

Finally, it seems like wireless routers are more expensive than
wireless access points.  I would think it would be the other way
around, more functionality in an access point than a router.

Oh I meant to say earlier, I set up my two laptops at a friends
house on his network, unfortunately he did a lot of the work so
I didn't really learn anything.
  

    
When I wrote the book (about a year ago now), 802.11g was just getting
started and there weren't any common standards for it yet. And no evaluation
hardware available. So I'm workign from what I've read, rather than my own
hands-on experience.

That said, I assume that configuring 802.11g isn't significantly different
from configuring 802.11b, at least as far as the users are concerned. If we
do a revised edition, I'll devote some space to 802.11a and g (and a lot
more attnetion to public hotspots).

I believe the 22 Mbps hardware uses a non-standard proprietary form of
compression to double the speed. From what I've seen, it works okay if it's
talking to an access point made by the same vendor as the network client,
but in a mixed network, it'll drop down to the slower rate.

Keep in mind that the slowest part of your network feed drives th speed of
the whole connection. So if your Internet feed is a DSL at, say, 650 kbps,
or even a T-1 at about 1 Mbps, you're not going to see any meaningful
difference between an 11 Mbps radio link and a 22 Mbps or faster link. It'll
make a difference on local file transfers, but not on your Internet service.

Routers cost more than access points because they're more complicated and
contain more parts. An access point is strictly the interface between WiFi
and Ethernet; a router adds some wired Ethernet links to the same box. I
don't know what the relative cost of goods and development costs are for the
two typess, but the router has to be more expensive to build. On the other
hand, a combined router/access point will be cheaper than two separate
boxes, because they share a power supply and other components. My experience
has been that the combined units on sale at Office Whatsis all seem to work
acceptably--Linksys, D-Link, Belden, Microsoft, whatever is on sale.
  

    
Wireless seems to be a fad for the fashionable currently with every
company from Starbucks to McDonalds thinking they can leverage their
franchise networks into something virtually profitable.

Is wireless being adopted that quickly?

I'm wondering if 802.11g is simply a warmup technology for something
greater. Is there anything on the horizon?

I'm also wondering when every elemetry school kid will have a laptop. It
seems that we aren't too far away from that - perhaps 5 years? Any
guesses?
  

    
Or would they more likely have a wireless Playstation or Gameboy?
  

    
I knew that 802.11g was not in place when you wrote your book,
and it specifically says it was about 802.11b, I was just wondering
if you had any comments about 802.11g.  Pointing out that the
limitation was the connection to the internet was helpful.

More info about hotspots would be helpful.

And, I mispoke.  Routers are *less* expensive than access points,
at least in the Tigerdirect catalog I was looking at.  Across the
board, all brands, usually about $10 or so.  That is what seemed
odd to me, it did seem that routers should cost more than an
access point.

Thanks for the info on 22Mbps.  I think I'll stick with 802.11g,
which probably is a warmup technology for something greater.  But
then, every technology seems to fall into that category.

I already have a wired router.  When I get a wireless router,
should I put it in front of the wired router, behind the wired
router, or just replace the wired router?  I guess it depends
partly on how many wired connections I need.  And I would need
to disable one of the DHCP servers, as I understand it.

Off to Office Depot, to see what is on sale
  

    
I went off looking for a simple access point that would act as a
wireless hub/bridge to my 10BaseT network that has the access router
and servers on it, and found that the AP+routers were cheaper than
non-routing APs.  My assumption is because that's the mass-market
product: most people seem to buy these for home networks and it's
their only router, and by default it's set up to get a single dynamic
address on the WAN side (PPP, DHCP, or PPPoE) and route to a NAT'ed
LAN where the router distributes DHCP leases.  

Anything other than that, including acting as a simple hub/bridge,
seems to be a specialty item, not intended for the mass home market,
and thus more expensive.   
  

    
Looking at the ads for laptop computers, it's my impression that a huge
proportion of new laptops are coming with internal WiFi. I assume they're
putting the antenna inside the clamshell around or behind the screen, which
implies that they'll have higher gain than a plug-in PCMCIA card. The
additional cost of the WiFi module (which plugs into the motherboard) is
trivial, but I'm wondering if this will mean that a lot of unsophisticated
users will be transmitting all the time. Seems like that will be an
unnecessary battery drain, and a minor source of RF pollution.

Anyway, at what? under $50 for a WiFi card and dropping, I'd guess that most
new portables will have either an internal or external WiFi unit within a
couple of years, or even sooner.

As for the access point vs router/switch question, a router makes sense in a
home network, because at least one machine will generally be within a foot
or two of the access point. So it's easier and cheaper to connect through a
$2 ethernet cable than with a $50 WiFi card. And you can use the wired link
to configure the access point.

>I'm also wondering when every elemetry school kid will have a laptop.

That's more a matter of educational philosophy than technology, isn't it?
Until middle school, I'd be concerned about security and durability. Can you
picture what a portable computer would look like after an eight-year-old has
been hauling it around for a few months? Seems like you could accomplish the
same thing with a file server without the danger of the kid leaving a $600
box on the bus. Or using it as second base.
  

    
It wouldn't be a $600 box anymore, most likely.  And you're right, might
as well buy a router/switch than an access point, but when what's
the point of access points (oops, pun not intended) if they cost more
than routers/switches?

In what case would you want *just* an access point?
  

    
> As for the access point vs router/switch question, a router makes
> sense in a home network, because at least one machine will generally 
> be within a foot or two of the access point.

I think it's more the case that you *need* a router if you are going
to use more than one machine, unless you have multiple public IP
addresses (less and less likely for redisentiall customers) or use a
soft router like Internet Connection Sharing in Windows or MacOS.  
So you have to have something to plug into the DSL or cable modem to
route your bits for you and do NAT and DHCP, which you won't typically
get (without charge) from the ISP at the lowest tier service.  

These days I don't think you can assume anything about the default
location of the machines...
  

    
slip
  

    
Uh, "residential". 

> In what case would you want *just* an access point?

I wanted it in order to have a nonrouting wireless hub, where Wi-Fi
clients would use public IP addresses from my block, thus obviating
the need for NAT (which makes some things difficult, like VPNs,
servers on static addresses that use multiple TCP ports, etc.).  
In other words just like a regular Ethernet hub (which does not figure
into addressing and routing). but wireless. 
  

    
I can see the telcos having a fit over WiFi. Voice over Ip will be
upsetting to them in terms of lost revenue.

However, looking at Blackout 2003, the wire telephone network is solid
while the WiFi and Cellular networks are tenuous.